Retrospective Theses and Dissertations
1962
Infiltration of water into soils as influenced by
antecedent moisture
Richard Ervin Green
Iowa State University
Follow this and additional works at: http://lib.dr.iastate.edu/rtd
Part of the Agriculture Commons
Recommended Citation
Green, Richard Ervin, "Infiltration of water into soils as influenced by antecedent moisture " (1962). Retrospective Theses and
Dissertations. Paper 2087.
This Dissertation is brought to you for free and open access by Digital Repository @ Iowa State University. It has been accepted for inclusion in
Retrospective Theses and Dissertations by an authorized administrator of Digital Repository @ Iowa State University. For more information, please
contact [email protected].
This dissertation has been
62-6495
microfilmed exactly as received
GREEN, Richard Ervin, 1931MFILTRATION OF WATER INTO SOILS AS
INFLUENCED BY ANTECEDENT MOISTURE.
Iowa State University of Science and Technology
Ph.D., 1962
Agriculture, general
University Microfilms, Inc., Ann Arbor, Michigan
INFILTRATION OF WATER INTO SOILS
AS INFLUENCED BY ANTECEDENT MOISTURE
by
Richard Ervin Green
A Dissertation Submitted to the
Graduate Faculty in Partial Fulfillment
The Requirements for the Degree of
DOCTOR OF PHILOSOPHY
Major Subject:
Soil Management
Approved:
Signature was redacted for privacy.
In Charge of Major Work
Signature was redacted for privacy.
Hea
Major Department
Signature was redacted for privacy.
Dear* of ^Gradua
illege
Iowa State University
Of Science and Technology
Ames, Iowa
1962
ii
TABLE OF CONTENTS
Page
INTRODUCTION
1
REVIEW OF LITERATURE
5
Influence of Antecedent Moisture on Infiltration
Development and Application of Moisture
Diffusion Theory
Measurement of Capillary Conductivity
Field Measurement • of Infiltration
Aggregate Stability and Surface Sealing
EXPERIMENTAL PROCEDURES
General Approach
Field Procedure
Laboratory Procedure
Calculation of Infiltration from
Conductivity Data
RESULTS AND DISCUSSION
Physical Properties of Soil Profiles
Aggregate Stability and Surface Sealing
Moisture Retention, Diffusivity,
and Capillary Conductivity
Field Infiltratiôn
Calculated Infiltration
5
9
25
29
30
34
34
35
42
53
60
60
61
71
82
100
SUMMARY AND CONCLUSIONS
123
LITERATURE CITED
128
ACKNOWLEDGMENTS
134
APPENDIX A
135
APPENDIX B
138
1
INTRODUCTION
Water infiltration into soils is a process of major
importance to agriculturalists and hydrologists and to the
general public.
Industrial and domestic demands on water
resources are increasing rapidly as the urban population con­
tinues to expand.
Increased crop production in agriculture
to meet growing population needs requires a more judicious use
of existing water sources both in irrigated and non-irrigated
regions.
Soil management research has as one of its principle
objectives the development of tillage and cropping practices
which are conducive to high rates of water infiltration.
The importance of water infiltration into soils has been
recognized for several decades, as is evidenced by the de­
tailed studies of Wollny in Germany as early as 1874 (Baver,
1938).
In the United States, extensive research on infiltra­
tion was undertaken during the 19301 s when soil and water
conservation became a matter of national concern.
Hy­
drologists, interested in applying the infiltration approach
to prediction of surface runoff from watersheds, required
quantitative estimates of water intake rates of soils over a
wide range of cover and soil conditions.
Infiltration rates"*"
"*"The term "infiltration rate" is used throughout this
dissertation in accordance with the usage recommended by the
Soil Science Society of America Committee on Terminology
(Richards, 1952). It is synonymous to "infiltration capacity"
2
of numerous soils were measured by a variety of methods, and
the effects of tillage, surface sealing by rainfall impact,
moisture content, and other important factors were noted by
many workers.
o
Antecedent soil moisture
which influence infiltration.
is among the major factors
Since unsaturated flow of water
in soils is closely related to the soil moisture content, the
used widely by hydrologiste, and is defined as the maximum
rate at which a soil in a given condition at a given time
can absorb water applied in excess at the surface, either as
rainfall or shallow impounded water. The infiltration rate
of a soil does not usually refer to a discrete value, but to
the graph of water entry rate versus elapsed time, since the
rate of entry generally decreases rapidly with time after in­
filtration begins. Thus, both the magnitude of intake rate
values and the shape of the infiltration rate versus time
curve are implicit in the term "infiltration rate". Since
the rate of water intake does approach a constant value with
time some investigators have used the "final" or "equilibrium"
infiltration rate to characterize a given soil or soil surface
condition with respect to its ability to absorb water, but
such a discrete value gives only limited information and is
not adequate for most applications. The frequent mention
herein of the "effect" of various factors on infiltration
refers to the way in which factors such as moisture content,
surface sealing, etc. influence water intake rates with time
during the period of infiltration.
n
The term "antecedent soil moisture" or "antecedent
moisture" will be used to refer to the soil moisture status
of the entire soil profile previous to the beginning of rain
or application of water to the soil. The word "antecedent",
meaning prior or going before in time, has been used widely
by hydrologists in connection with soil moisture effects on
runoff and infiltration. The term "moisture effect" will be
used occasionally for brevity, and refers to the effect of
the antecedent moisture status of the soil on subsequent in­
filtration. "Initial moisture", as used herein, refers to the
moisture content of a discrete soil layer or soil sample.
3
infiltration rate is dynamically influenced by the moisture
distribution within the soil profile, both at the beginning
of infiltration and at later times.
The importance of the
moisture effect on infiltration has been widely recognized
and measured qualitatively, but quantitative estimates of the
moisture effect over a broad range of antecedent moisture
conditions have been difficult to obtain.
Field infiltration
studies, using sprinkling infiltrometers, are being conducted
to determine the effect of tillage and other soil management
practices on infiltration and to characterize different soils
with respect to infiltration.
Such research is hindered in
that results from field measurements of infiltration on a
given soil with given surface conditions may vary considerably,
depending on the moisture status of the soil profile.
Thus,
it is extremely important to both hydrologie and soil manage­
ment research that procedures be developed to account for the
moisture effect on infiltration.
Field measurements are ex­
pensive and generally limited to a few antecedent moisture
conditions.
Statistical correlation of infiltration with
various indices of soil moisture content have provided only
limited information on the moisture effect.
Recent work by Philip (1957f) suggests that the moisture
effect on infiltration should be estimable by the application
of moisture diffusion theory, utilizing laboratory measure­
ments of capillary conductivity and soil moisture retention.
4
However, until recently, the only available method of solving
the moisture flow equation assumed an isotropic soil and
required that the antecedent moisture content be uniform
with depth, conditions which are seldom met in the field.
A recently developed numerical procedure (Hanks and Bowers,
1962a) allows solution of the flow equation for layered soils
with non-uniform antecedent moisture profiles, thus providing
a means of applying theoretical developments to actual field
problems.
The purpose of the present study was to investigate
prediction of infiltration rates as influenced by antecedent
moisture, making use of Hanks and Bowers' numerical procedure
and employing recent improvements in the laboratory measurement of capillary conductivity.
Field measurements of in­
filtration on two soils under a variety of antecedent moisture
and surface conditions provided a means of checking the
validity of theory calculations and gave a measure of the
moisture effect on infiltration relative to the effects of
tillage, surface sealing, and soil morphological differences.
Attention was also given to the effect of soil moisture con­
tent on aggregate stability and surface sealing.
O
The terms "capillary conductivity" and "capillary diffusivity" are not synonymous, but may be used interchangeably
in the practical sense in that they both refer to "the physi­
cal property relating to the readiness with which unsaturated
soils transmit water" (Richards, 1952, p. 87). The distinction
between the two terms will be clarified in later discussion.
D
REVIEW OF LITERATURE
Influence of Antecedent Moisture on Infiltration
Norton (1933), one of the first to recognize the impor­
tance of infiltration in the hydrologie cycle, observed that
in the field the maximum water intake rate was obtained during
short storms following dry periods, and minimum rates occurred
during prolonged wet periods.
Thus, he recognized an ante­
cedent moisture effect on infiltration, but later expressed
the opinion that the infiltration process was controlled
primarily by conditions at the soil surface (Morton, 1940).
Emphasis by Duley (1939) on the effect of surface sealing in
decreasing the rate of infiltration supported Morton's view­
point.
It remained for Bodman and Colman (1944) to show by
careful laboratory experiments that the decrease in infiltra­
tion rate with time was proportional to the decrease in
potential gradient within the "transmission" zone during
infiltration.
The combined work of the forementioned re­
searchers emphasizes that water infiltration into soils is a'
dynamic process in which both the moisture potentials within
the soil and soil structural conditions at the surface may
change strikingly during the early stage of infiltration,
causing a reduction in infiltration rate with elapsed time.
Infiltration is closely related to the pore space
6
available for storage of infiltrating water (Hansen, 1955).
On some soils the total pore space may be reduced when soil
colloids swell upon wetting.
Browning (1939) showed that if
the volume change of a soil (especially subsoils) exceeded 20
percent upon wetting, then the infiltration rate was reduced
to extremely low values.
The importance of the antecedent moisture effect on in­
filtration has been emphasized by a number of researchers in
both laboratory and field studies.
Neal (1938), in a labo­
ratory study of the effects of degree of slope, surface con­
dition, and rainfall intensity on infiltration, found that
initial soil moisture content had a greater effect on the
rate of infiltration during the first 20 minutes of simulated
rainfall than any other factor.
Other workers (Free et al.,
1940; Arend and Morton, 1943) obtained similar results in
field measurements with infiltrometers.
Brakensiek and
Frevert (1961) emphasized the need of accounting for ante­
cedent moisture when using infiltrometer data to estimate
runoff on actual watersheds.
Jamison and Thornton (1961) in
an analysis of 12 years' rainfall-runoff data from runoff
plots and field terraces under various surface and cover
conditions found that antecedent moisture was the dominant
factor determining infiltration rates.
Similar results were
obtained by Thames and Ursic (1960) on small watersheds in
northern Mississippi where surface runoff was found to be
7
strongly correlated with available storage in the upper six
inches of soil.
Using runoff plot and rainfall data, Bertoni
et al. (1958) calculated infiltration rates on Marshall silt
loam and attempted to determine the effect of wetness of the
soil, slope length, and season of the year on infiltration
rate.
An antecedent precipitation index (API) was correlated
with various infiltration indices as a measure of the effect
of soil wetness on infiltration.
Antecedent precipitation
was highly correlated (negatively) with cumulative infiltra­
tion but very lowly correlated with infiltration rate.
Using
a similar approach, Ligon and Johnson (1960) found the API to
be unrelated to either the initial or final infiltration rate.
These runoff plot results indicate the difficulty of defining
the extent of causal relationships in field infiltration
studies in which a number of factors vary beyond the control
of the experimenter.
A few researchers have sought to evaluate the effect of
antecedent moisture on infiltration in field experiments by
statistical correlation of cumulative infiltration for a
given time with the moisture content or "available storage"
in the surface 6 or 12 inches of soil.
Tisdall (1951),
Reinhart and Taylor (1954), and Ayers and Wikramanayoke
(1958) used similar procedures and obtained significant
correlations between antecedent moisture and cumulative
infiltration.
Linear regression equations were developed
8
which could be used to predict cumulative infiltration in a
given time as a function of the initial moisture content.
This approach is useful in irrigation work where it is de­
sirable to know how much water will enter the soil during a
given period of application.
However, the prediction equa­
tions obtained are of limited value since they are applicable
to only those time periods for which they were derived, and
in addition, they provide little information on the rate of
infiltration with time.
Other workers have attempted to predict infiltration by
accounting for both "available storage" and the rate of water
movement through field cores.
Schiff and Dreibelbis (1949)
related infiltration rates to antecedent soil moisture by
measuring the rate of water movement through field cores at
different initial moisture contents.
Their recognition of
the influence of antecedent moisture on infiltration and
their attempt to predict the moisture effect utilizing labo­
ratory measurements are noteworthy.
Diebold (1951), in a
similar approach, found that when a rate prediction equation
included a saturated conductivity factor and a water storage
capacity factor, the variables accounted for 55 percent of
the variation in field infiltration rates, as compared to 44
percent without the water storage factor.
Despite the shortcomings of the above methods, they
represent attempts to incorporate into one prediction equation
9
two factors which are now known to influence infiltration
rates, the rate of moisture flow in the soil and the volume
of empty pore space which can be filled by incoming water.
It is doubtful that adequate prediction of infiltration rates
or of the moisture effect on infiltration can be accomplished
without defining more precisely the functional relationship
of infiltration to the soil physical properties governing
moisture flow.
Development and Application
of Moisture Diffusion Theory
In addition to the voluminous empirical evidence of the
effect of antecedent moisture on infiltration in field and
laboratory studies, developments in the theory of unsaturated
moisture flow have also emphasized the moisture dependent
nature of water movement in porous materials.
Considerable
progress has been made in describing moisture flow mathemati­
cally utilizing diffusion theory and in identifying the physi­
cal parameters which govern moisture flow.
The purpose of
this section is to review briefly the development of moisture
diffusion theory, including some simple derivations of flow
equations.
Such review appears worthwhile to provide adequate
background for later discussion and to show the pertinence of
theoretical developments to practical infiltration problems.
10
Buckingham (1907), in his studies of capillary action in
soils, proposed that moisture flow in unsaturated soils was
roughly analogous to heat flow through an insulated metal bar
or the flow of electrical current through a wire.
Thus he
proposed that capillary flow of water could be described by
an equation similar to that describing heat or electrical
flow, that is Q = Xs where Q is the current density or flow
velocity per unit area, S is the gradient in the attractive
force, and X denotes the capillary conductivity of the soil.
His capillary rise experiments demonstrated (Buckingham, 1907,
p. 27) that water flows "from places of high to places of low
water content, just as heat flows from high to low temperature,
or electricity from high to low potential". "With regard to
moisture flow, Buckingham states, "the driving force or cause
of the current is the different attraction for water of two
portions of the soil that are not equally moist", and "the
conductance of the soil is the facility with which it allows
water to flow through it, or its power to conduct water".
Thus, by analogy with the known laws of heat and electrical
flow, Buckingham expressed unsaturated moisture flow in terms
of the capillary conductivity and the gradient in capillary
potential.
In addition to this significant advance in de­
fining the dynamics of unsaturated moisture flow, Buckingham
observed that both the capillary potential, t}>, and the capil­
lary conductivity, K, were not constants for a given material
11
but were largely dependent on soil moisture content, 9.
Thus
Buckingham sought to investigate the functional relationships
i|)(6) and K(9).
Two ideas arising from Buckingham's work
which are of special interest in the present study are:
(a)
Moisture flow in unsaturated soils is governed by
physical properties of the soil-water system and
should be amenable to description by a mathematical
equation of physical significance.
(b)
The two physical properties relevant to moisture
flow, capillary conductivity and capillary poten­
tial, are both moisture dependent variables,
emphasizing the inherent importance of soil
moisture content in all processes involving un­
saturated moisture flow, such as infiltration.
The flow of water through uniform saturated soils had
been adequately described before the time of Buckingham by
Darcy's law, which is analogous to Fourier's law, Ohm's law,
and Fick's law, and may be expressed for one dimension by
the equation
where v is the quantity of water conducted per unit time
across unit area in the x direction, K is the hydraulic con­
ductivity and ai/3x is the potential gradient in the x direc­
tion.
In the saturated case, the proportionality factor K is
a constant for:a given pore geometry and water viscosity, and
thé potential gradient results from a positive head differenc
12
between two points.
In unsaturated flow, on the other hand,
the conductivity depends on the moisture content (i.e. it is
not constant) and the potential gradient results from both
capillary and gravitational forces.
Richards (1931) assumed
the applicability of Darcy's law to unsaturated flow, and
combining it with the equation of continuity, derived a dif­
ferential equation for capillary flow with the capillary
potential, i|», as the dependent variable.
Other workers
(Gardner and Widtsoe, 1921; Childs, 1936a, b) used the dif­
fusion theory approach implicit in the concepts proposed by
Buckingham, but assumed a constant diffusion coefficient.
This approach was later shown experimentally to be inap­
propriate to unsaturated flow, owing to the moisture dependent
diffusivity (Kirkham and Feng, 1949).
For the purpose of definition, it should be noted that
the diffusion coefficient or "diffusivity", D, arises as a
proportionality constant in a diffusion equation of the type
used in Fick's first law to express the movement of solute
down a concentration gradient, i.e.
q = -D^
(2)
where q is the mass of solute passing in the x direction
through unit area per unit time, and c is the concentration
(mass per unit volume).
The D in this case has the dimensions
13
(length)^ per unit time (Hodgman, 1960)
The D also occurs in Fick's second law which is derived
by combining the first law with the equation of continuity.
The continuity requirement in this case is given by
<3>
*t = -£
which states, essentially, that the change in concentration
of solute (gms.) in a given unit volume of solution equals the
amount coming in minus the amount going out.
The equation
which results from combining Equations (2) and (3) is
(4
=
>
Since in this case the diffusivity, D, is constant, the equa­
tion for one dimension may be written in the form
t5 >
» • 4Equation (5) is essentially the same equation derived by
Childs (1936a) to describe moisture flow, assuming constant
diffusivity.
The use of the term "moisture diffusion" to
describe the flow of water in unsaturated soils is not really
appropriate since it is not a diffusion phenomenon, but
rather a case of mass flow through films.
However, because
14
capillary flow can be described by an equation such as (4)
with D dependent on the moisture content, 9, the D(9)
function
X
has been called the diffusivity.
Klute (1952) derived the partial differential equation
describing capillary flow in porous materials and solved it
for the case of horizontal flow.
The derivation of the mois­
ture flow equation is similar to the derivation of (4), and
is instructive to follow briefly.
In the moisture flow case,
Darcy's law, (1), is combined with the continuity requirement,
i.e.
<6>
#=
to obtain
If =
where K is now the unsaturated hydraulic conductivity.
Equa­
tion (7) gives the change in moisture content with time, when
x is held constant, as a function of the conductivity and the
potential gradient.
In unsaturated flow the total potential,
$, represents the sum of the capillary potential, t|j , and the
Since for unsaturated moisture flow all variables are
considered to be functions of the moisture content, it must
be kept in mind that D = D(9), K = K(9) and tJj = i|>(9). For
brevity the symbols D, K and
will be used.
15
gravitational potential.
called "tension",
The capillary potential (often
11 suction"
or "negative pressure") may be
expressed as an equivalent height of a water column.
If the
soil at a height z above a reference plane has a capillary
potential, t|j, then the gravitational potential at that height
is z, and the total potential is given by
= i|j + z.
(8)
If we let z = -x in (8) so that for the infiltration case x
will be positive downward in the vertical dimension, and then
replace Q in (7) by 3> = i|) - x, we get ^
].
Taking the gravity term outside the brackets, we have
S
(
?
9
>
The gravity term, a K / a x , would equal zero in the case of
horizontal flow.
Assuming a single-valued relation between
t|j and 0 we have by the chain rule,
= KH H - DH
uo)
where D = K-^, and is called the diffusivity, with dimensions
(length)^per unit time.
Substitution of 0—^ for K-^ in (9)
results in Equation (11), a partial differential equation
16
capable of describing the vertical movement of water in a
semi-infinite column of uniform soil.
H
=
(ID
The similarity of Equation (11) without the gravity term to
Equation (4) explains why Equation (11) is called a "moisture
diffusion equation".
The relationship of the diffusivity, D,
to the conductivity, K, that is, D =
should be noted,
since both terms are used frequently to describe the moisture
conducting properties of unsaturated soils.
Equation (11), subject to the conditions
t = 0, x>0, 0 = Q
"
(12)
x = 0, t > 0, 0 = 6
Q
describes water infiltration (Philip, 1957b), where
is the
antecedent moisture content (constant for the entire profile)
0 is the moisture content for t|) = 0, i.e. at saturation, x i
Q
the depth, positive downward, and t is time.
The moisture
3
content is expressed on the volume basis, e.g. cm. water per
cm.^ soil.
Note that the condition t = 0, x > 0, 0 = 9fi
requires a uniform moisture content with depth at the be­
ginning of infiltration.
Philip (1957a) has developed a numerical procedure for
17
solving Equation (11) subject to (12) and has investigated on
theoretical grounds a number of problems pertinent to infil­
tration.
Philip's work, given in a series of seven papers
entitled "The Theory of Infiltration" (Philip, 1957b, c, d,
e, f; 1958a, b), represents an extremely important application
of moisture diffusion theory, and is the first comprehensive
analytical study of water infiltration.
Philip has shown that the solution of Equation (11)
subject to (12) can be expressed in a power series of time,
the coefficients of which are functions of the moisture con­
tent and are found as solutions to a series of ordinary dif­
ferential equations by numerical methods.
The solution to
(11) is of the form
x = at^ + J3t + yt3//^ + ... + fm(Q)tm//2 + ...
with a = a(9), |3 = {3(9), etc.
(13)
For the range of t and D- and
K-functions of interest to soil scientists, Series (13) con­
verges so rapidly that only a few terms are needed.
Once the
coefficients have been found by numerical procedures, Equation
(13) can be used to determine the moisture profile with time,
cumulative infiltration, and infiltration rate.
Philip
(1957b, p. 352) states:
Since the total change of moisture content in the
semi-infinite column equals the difference between
18
the time integral of the flux at x = 0 and at
infinity, then
90
l =
xde + Knt
(14)
®n
where i denotes the cumulative infiltration, K is
the conductivity at the initial moisture conteRt,
9 , and t is time.
This means, in essence, that the moisture accumulated during
a given time interval is the difference between the quantity
entering at the surface and the quantity leaving the bottom.
However, the derivation of (14) is not immediately obvious.
Figure la illustrates graphically the derivation of Equation
(14).
The graph labeled t = 0 gives the antecedent moisture
content (dashed line) with depth prior to application of
water.
Note that this initial condition satisfies the re­
quirements of (12).
After infiltration has occurred for a
periiod t^, the moisture profile is represented by the graph
labeled t = t^.
The shaded area above the curve equals the
volume of moisture which has accumulated during the period
t = 0 to t = tp and represents the first term in (14)
90
xd9.
6n
Corresponding to 9
Q
is a conductivity K and corresponding to
Q
the initial moisture, 0 , is a conductivity Kn.
During the
Figure la.
Schematic diagram to illustrate the deriva­
tion of Philip's expression for cumulative
infiltration,
6.
l -
xdG +
e
Figure lb.
n
Predicted influence of initial moisture con­
tent, 9n, on infiltration rate ; numbers cor­
responding to each curve denote values of 0
in percent (taken from Philip, 1957f,
p. 331)
INFILTRATION RATE
H
-.0
PO
CJl
en
ro
o
OJ
CD
CD
CD
21
period of infiltration there will be a flux of Kn cm. per
hour at the point x^ in the profile, so that the total amount
of water passing by x^ in t^ hours is K^t^ cm., i.e. the last
term in (14).
For low values of ©^ (dry soil) Kn will be low,
and the K t term of (14) will be negligible.
On the other
hand, when 9n is high (wet soil) Kn will be higher and the
K t term may account for a sizeable fraction of the cumulative
n
infiltration.
integral
However, with ©n high, the value of the
r®0
xdG can be expected to be lower for a given
8n
period than when Qn is low, so that cumulative infiltration
is generally lower for the wet case, in accord with field
results (Reinhart and Taylor, 1954).
Since the x in (14) is defined in terms of a power
series of time in (13), when the indicated integration of
(14) is accomplished, cumulative infiltration is given as a
power series of time,
i =
90
r 90
ad© + t [
P + Kn] +
9n
(15)
9n
90
n 60
ad© is denoted by S and [
pd© + K ] is denoted by
If
9n
9n
A, and the Series (15) is truncated after the second term,
22
then we have
(16)
i = St^ + At.
Differentiation of (16) with respect to t results in
(17)
Vq = %St ^ + A,
where V is the infiltration rate witji time expressed in terms
Q
of a simple algebraic equation.
This equation, based on
physical theory, has been shown to compare favorably with
empirical equations in common usage (Philip, 1957e; Watson,
1959).
The reliability of Philip's solution has been demon­
strated in both the laboratory and field.
Youngs (1957) con­
ducted laboratory infiltration measurements on columns of dry
porous materials and obtained results which agreed well with
those computed by the analytical solution.
Nielsen et al.
(1960) obtained reasonable agreement between calculated mois­
ture profiles and those measured on two. relatively uniform
loess soils.
Discrepancies between measured and calculated
results were explained on the basis of non-uniformity of the
soil profile.
Supported by both laboratory and field results,
the analytical solution can be considered essentially valid,
and the further analysis of the infiltration process by
23
Philip accepted with some confidence.
Of particular interest to the present study is the fifth
paper of Philip's series (Philip, 1957f) which considers the
influence of initial moisture content on infiltration.
Figure lb (from Philip, 1957f, p. 331) shows the influence
of soil moisture content on infiltration rate curves computed
for Yolo light clay.
Two points are emphasized by Philip
with reference to Figure lb:
first, the marked dependence of
infiltration rate on ©n at small times, and second, the fact
that as time increases, @n has less and less effect on the
intake rate.
Although the curves in Figure lb have not
reached a common rate of intake for the largest time shown,
the theory predicts that all curves will be asymptotic to the
line
K
Q
(conductivity at saturation) as time increases with­
out limit (Philip, 1957c).
Although increased initial moisture has a depressing
effect upon the infiltration rate, it increases the rate of
advance of the wet front.
This seemingly contradictory re­
sult is predicted by theory and has been shown experimentally
by Colman and Bodman (1944).
In one paper, Philip (1958a) considered the effect of
water depth over the soil on infiltration.
From the analysis
he proposed that for a given water depth and time, the depth
of the saturated zone increases as 9n increases.
In addition,
the rate of infiltration is increased with increasing water
24
depths over the soil, the depth effect being most profound on
initially wet soils.
Philip (1957f) emphasized that the infiltration process
involves complex interactions between initial moisture con­
tent, moisture gradients in the profile, and the moisture
dependence of the diffusivity and conductivity functions,
making difficult the separation of clear-cut causal rela­
tionships which might enable one to explain, for example, the
initial moisture effect.
Among the more complex problems not treated in the
analysis by Philip are the following:
(a)
infiltration into layered soils, or soils
which are non-homogeneous in the vertical
dimension, and
(b)
infiltration into soils, either uniform or
layered, with non-uniform moisture contents
with depth.
A recently developed numerical method by Hanks and Bowers
(1962a) makes possible the solution of a finite difference
form of Equation (7) for the initial and boundary conditions
arising in the problems mentioned above.
The method requires,
in addition to a knowledge of the boundary conditions of the
specific problem,
(a) ' a known relationship between moisture content and
pressure head, i.e. the i|)(©) relation, and
(b)
a known relationship between moisture content and
moisture diffusivity, i.e. the D(Q) relation.
25
The method does not require that
(a)
the soil be homogeneous in the vertical dimension,
(b)
the soil depth be semi-infinite, or
(c)
the initial moisture content be uniform.
Hanks and Bowers (1962a) compared results from their numerical
method with results from Philip's method for the case of
horizontal flow, and obtained excellent agreement.
Computa­
tions for infiltration into layered soils gave reasonable re­
sults, but for this case there were no other numerical methods
for comparison.
It appears, then, that this method of solu­
tion, making use of finite difference equations and a pro­
cedure adapted to high speed computers, may extend the
application of diffusion theory to infiltration problems
which are beyond the scope of previous approaches.
Without
the "homogeneous soil profile" and "uniform moisture profile"
restrictions, the method may be useful to predict infiltration
rates for the numerous boundary conditions existing in the
field.
Measurement of Capillary Conductivity
The use of the moisture diffusion equation for a given
soil requires knowledge of the functions K(9) and t|)(9).
The
i|j(9) relation is known to exhibit hysteresis when a desorption
26
process is followed by absorption.
However , ;j>(9) can be
determined for both drying and wetting processes using pres­
sure plate apparatus, and then the appropriate part of the
hysteresis loop used for the phenomenon being studied.
The
measurement of K(9) or D(9) (one can be obtained from the
other if the t|)(9) relation is known) is difficult, and all
current methods have some limitations.
»
The first published laboratory method of determining the
capillary conductivity of a soil (Richards, 1931) involved
applying a differential water suction to the two ends of a
soil sample and measuring all terms in Darcy's equation except
the unknown K.
This method has been widely used and is espe­
cially well adapted for use with "disturbed" soil samples and
for suctions less than 200 cm. of water.
A new apparatus for
use with this method has been described by Nielsen and Biggar
(1961).
Another method, similar to that of Richards and used
by Nielsen et al. (I960) allows evaporation from one end of a
soil core to establish the pressure potential between the two
ends of the core.
In both of these methods the potential
difference is measured with small tensiometers, and the con­
ductivity is calculated under steady-state flow conditions.
Owing to the low air-entry values of tensiometers and mem­
branes generally used in these methods, they are limited to
measurement of K at suctions less than 200 cm. water.
Ceramic
plates can be used, but cause a significant pressure drop
27
across the plate, and require long periods of time for the
sample to reach equilibrium.
Bruce and Klute (1956) developed a method, based on dif­
fusion theory, by which the D(9) function is calculated from
the moisture content distribution curve for a soil core which
has absorbed water at one end.
The method yields only ap­
proximate values of diffusivity, especially at high moisture
contents, but provides a rapid and relatively easy means of
obtaining the diffusivity for disturbed soil samples.
This
method has been used by Hanks and Bowers (1962a), and the D
values obtained yielded reasonable results when applied to
various moisture flow problems.
One limitation of the procedures discussed above is the
necessity of measuring the o|)(0) function on a separate sample
from the one on which K or D is measured.
Gardner (1956)
developed a method for determining capillary conductivity
using measurements of the transient outflow from a soil
sample in a pressure plate apparatus.
By measuring the out­
flow as a function of time and using the solution of the
moisture flow equation, the capillary conductivity can be
calculated for each suction interval.
It is assumed that for
small pressure increments (a) the capillary conductivity is
approximately constant, and (b) the relation between the
water content and suction is linear.
Both of these condi­
tions appeared to be met if the pressure increment was
28
sufficiently small.
Gardner showed that when outflow mea­
surements were made using conventional ceramic plates the
boundary condition requiring the lower boundary of the soil
to be atmospheric at all times was not met for pressures less
than 100 millibars.
The resistance of the plate to water
flow (plate impedance) was not negligible compared to the
resistance to flow within the soil.
Although this procedure
does not allow one to measure capillary conductivity in the
0-100 millibar pressure range, it makes possible the measure­
ment of K at high pressures, and allows the simultaneous
determination of the K(8) and x|)(6) functions.
A refinement of Gardner's method was worked out by
Miller and Elrick (1958) who accounted for membrane impedance
in their analysis.
The major limitation of the Miller and
Elrick method lies in the uncertainty of determining the
membrane impedance.
In addition, unless good contact is
established between the soil and the porous plate, an unknown
contact impedance exists.
Kunze and Kirkham (1962) have modified Miller and
Elrick's method to simplify calculation of capillary con­
ductivity from pressure outflow data.
In addition to reducing
the computations, the method has provided two experimental
advantages:
(a)
It makes use of early outflow when, the authors
suggest, the diffusivity is essentially constant
29
owing to negligible moisture content changes re­
sulting from small outflow volumes. Also, the plot
of time versus outflow is very sensitive for the
early period.
(b)
The plate impedance need not be measured since it
is automatically accounted for in the reduction of
the data to a conductivity value.
The outflow procedures discussed above are well adapted
to studies involving the use of field cores.
Since field
cores cannot be duplicated exactly, the concurrent determina­
tion of K(0) and i|j(0) is a definite advantage.
Field Measurement of Infiltration
Water infiltration rates have been measured or estimated
by several methods including sprinkling infiltrometers, ring
infiltrometers, and hydrograph analysis on runoff plots and
small watersheds.
In many areas, methods making use of
natural rainfall, such as runoff plots and watersheds, re­
quire a number of years to obtain reliable data owing to the
low frequency of suitable runoff producing storms.
In addi­
tion, the number of variables, especially in the case of
watershed hydrograph analysis, make it difficult to obtain
clear-cut relationships between infiltration and causal fac­
tors.
Therefore, sprinkling infiltrometers and both single
and multiple ring infiltrometers have been used extensively.
In general, infiltration rates from ring infiltrometers are
30
higher and more variable than those obtained with sprinkling
infiltrometers.
In addition, sprinkling infiltrometers more
closely simulate natural rainfall allowing an estimate of the
effect of surface sealing by rainfall impact.
A new portable
infiltrometer, designed by Bertrand and Parr (1961) is being
used currently in most of the North Central States in a co­
operative attempt to standardize and jointly improve infiltra­
tion measurement equipment.
The "Purdue Sprinkling Infil­
trometer" , as it is called, can be operated by two men, and
is a substantial improvement over previously available infiltrometers.
Parr and Bertrand (1960) have published a comprehensive
review of infiltration measurement methods.
Aggregate Stability and Surface Sealing
The structural condition of the immediate soil surface
is known to influence infiltration rates.
On cultivated
soils, surface sealing due to rainfall impact may be the most
important factor controlling infiltration since the surface
seal at the surface often becomes the least permeable plane
in the profile (Duley, 1939; Ellison, 1945).
Mclntyre (1958),
by microscopic study of thin sections of soil crusts and mea­
surements of crust permeability, identified a thin compact
seal 0.1 mm. thick over a washed-in layer, the permeability
31
-7
-6
of these two layers being 5x10
and 5x10
cm. per sec.,
respectively.
The underlying soil had a permeability of 10
cm. per sec., emphasizing the restricting influence of surfac
sealing on water infiltration.
The sealing of soil surfaces depends to a great extent
on the stability of soil aggregates, i.e. their resistance to
slaking and breakdown when subjected to rainfall impact.
The
subject of aggregate stability is complex, and will not be
reviewed here in detail.
However, it is of interest in the
present study to give attention to the effect of initial mois
ture content on aggregate stability during a rain.
Yoder (1936) hypothesized that slaking of aggregates re­
sulted when capillary absorption of water into the aggregate
compresses entrapped air until the pressure developed finally
disrupts the aggregate.
Robinson and Page (1951) found that
large aggregates were more subject to slaking than small ones
and concluded that pressure build-up is more likely on large
aggregates.
If compression of air within an aggregate upon wetting
does occur, it seems logical that aggregate breakdown by this
means would be inversely related to initial moisture content
since more air can be entrapped in dry aggregates than in
those which are initially wet.
However, if the cementing
material binding the primary particles is weakened with in­
creasing moisture content, an opposite effect of moisture
32
content could result.
Cernuda et al. (1954) studied the
influence of initial soil moisture on the resistance of
macro-aggregates to slaking and to water-drop impact.
Aggre­
gates of size 1/4 to 1/2 inch in diameter were wetted under
vacuum and then subjected to tension to achieve several
initial moisture contents.
Upon subjection to slaking in
water, followed by water-drop impact, initially dry aggre­
gates were destroyed more easily than wet aggregates on all
soils studied.
The authors concluded that the disruptive
force of entrapped air was a more important cause of aggre­
gate breakdown than the weakening of cementing materials upon
wetting.
However, completely saturated soil aggregates were
•more easily destroyed by falling drops than aggregates at
moisture contents corresponding to low tensions, apparently
reflecting the cohesive action of interfacial water films in
the unsaturated aggregates.
They also noted that aggregate
stability was influenced less by initial moisture content on
coarse-textured soils than on fine-textured soils.
Evans (1954) found that air-dry soil which had been
moistened to the moisture equivalent 24 hours before wet
sieving had greater water stability than soil which had been
moistened only 5 minutes before analysis.
He concluded that
increased bonding associated with the longer wetting period
resulted from swelling and the resulting elimination of small
cracks in the aggregates, allowing continuous water bonds
33
across planes of weakness.
In summary, the antecedent moisture content of the soil
affects both the dynamics of moisture movement in the soil
profile and the stability of soil aggregates at the surface,
the former effect being quite well defined by moisture dif­
fusion theory, while the latter effect is not so well under­
stood.
34
EXPERIMENTAL PROCEDURES
General Approach
The effect of antecedent soil moisture on infiltration
was studied from two aspects:
(a)
the moisture effect on infiltration when the soil
surface was not subject to any major structural
changes during the infiltration period, and
(b)
the effect of initial moisture content on aggre­
gate breakdown and surface sealing by rainfall
impact.
The first aspect represents the major portion of the study
while the second is supplementary.
The general approach in
(a) involves prediction of infiltration rates for different
antecedent moisture conditions using laboratory measurements
of both capillary conductivity and moisture retention on field
cores.
The conductivity, K(9), and moisture retention,
relations for a given soil are used in the diffusion equation
for unsaturated flow, (11), and together with the stated
boundary conditions, specify the nature of flow in a given
infiltration problem.
In this way, infiltration rates for
various antecedent moisture conditions were predicted.
To
evaluate the accuracy of the predicted results, field mea­
surements of infiltration rate were made at the soil sampling
site.
In addition, infiltration rates were measured for dif­
35
ferent soil surface conditions in order to evaluate the
magnitude of the moisture effect relative to the effects of
tillage and surface sealing.
Laboratory measurements of
other physical properties of the soil profile were taken in
order to characterize the soils and determine the nature and
extent of horizon differentiation.
The moisture effect on
surface sealing (part (b) above) was studied in the labo­
ratory using a rainfall simulator.
Field Procedure
Soils
Two loess derived soils were chosen for the study: the
Ida silt loam, at the Western Iowa Experimental Farm near
Castana, Iowa, and the Grundy silty clay loam, at the Grundy
Shelby Experimental Farm near Beaconsfield, Iowa.
The Ida
soil, a Regosol with a very uniform textural profile, was
selected for the study of the antecedent moisture effect on
infiltration.
Earlier water movement studies by Nielsen et
al. (1959) indicated that the Ida was very homogeneous with
depth, except for a region at about 75 cm. depth where the
clay content was 3 percent higher than in the rest of the
profile.
A uniform soil profile was desired since moisture
flow theory is simpler for a uniform than for a layered soil
The Grundy soil, a well developed Brunizem, contains much
36
more organic matter in the surface than the Ida and has a
significant clay accumulation in the B horizon.
Infiltration
measurements on the Grundy provided an evaluation of surface
condition effects on a soil which differs strikingly from the
Ida in profile development.
Thus, the antecedent moisture
study outlined above in part (a) was done on the Ida, while
surface sealing was studied on both the Ida and Grundy soils.
Detailed profile descriptions taken at the sites of the field
measurements are given in Appendix A.
Antecedent moisture and soil surface conditions
To evaluate the effect of antecedent moisture on in­
filtration in the field, successive infiltration measurements
were made on bromegrass sod at three antecedent moisture
levels.
The terms "dry", "moist", and "wet" refer to the
relative moisture content of the 0-30 cm. soil depth for the
three antecedent moisture conditions.
The existing field
moisture at the beginning of the experiment was called the
"moist" condition, while the "wet" condition referred to the
moisture profile attained two days after the initial infiltra­
tion measurement, i.e. approximately the field capacity.
The
"dry" condition was achieved by evapo-transpiration from the
grass sod for a ten day period after the "wet" run.
The
bromegrass sod provided stability to the soil surface, and
thereby prevented the structural changes that would have
37
occurred with successive water applications on a tilled sur­
face.
In addition, a three-layer screen was placed about
four inches above the plot during the infiltration measure­
ment to prevent surface sealing due to water-drop impact on
the small areas of surface not covered with vegetation.
Laboratory trials showed that three layers of window screen,
2 to 3 cm. apart, effectively dissipated the energy of falling
water drops.
Infiltration rate as influenced by tillage was measured
on plots which were tilled to a depth of about 15 cm. and
covered with the three-layer screen during spray application.
All vegetation was removed from the surface prior to tillage.
Tillage by rototiller was chosen to provide a uniform tilled
surface layer that could be easily sampled to obtain uniform
soil cores for laboratory measurements of moisture retention
and conductivity.
On the Ida soil the rototiller loosened
3
the soil from an initial bulk density of 1.3 g. per cm. to
3
0.75 g. per cm. .
Owing to mechanical failure of the roto­
tiller, the Grundy soil was spaded to a depth of 15 cm. and
hoed to break up large aggregates.
On both the Ida and
Grundy the soil surface was raked smooth, the final raking
being down-slope to minimize surface retention during water
application.
The slope on both soils was about 6 percent.
The rototilling of the Ida soil left the surface 15 cm. in a
much more homogeneous condition than did the hand tillage of
38
the Grundy soil.
However, the soil tilled with the roto­
tiller was so unstable that it consolidated during water
application and developed large surface cracks shortly after
the cessation of "rainfall".
This change in surface condi­
tion eliminated the possibility of predicting infiltration
rates on the tilled soil, since laboratory measurements
could be made only on relatively stable soil cores.
A measure of the effect of surface sealing by rainfall
impact was obtained by a comparison of infiltration rates on
similarly tilled screen covered and bare plots.
Preliminary
work on a Clarion loam indicated that although there was some
aggregate slaking and surface sealing on the screen covered
soil surface, rainfall impact on a bare surface greatly in­
tensified aggregate breakdown and sealing.
At the beginning of the experiment the entire experimen­
tal area was covered with polyethylene film to maintain
similar moisture conditions on all plots during the 10-day pe­
riod needed to complete infiltration measurements.
Thus, the
moisture level of the tilled soil was equivalent to the
"moist" condition of the grass plots.
A randomized complete block design was used with three
replications on the Ida and two on the Grundy.
Infiltration
measurements were obtained on the Ida in June, 1961, and on
the Grundy in July, 1961.
Although measurements for two
antecedent moisture levels had been planned for the Grundy,
39
persistent rainy weather in July prevented more measurements.
Infiltration
Infiltration rates were measured with the Purdue
Sprinkling Infiltrometer which has been described in detail
by Bertrand and Parr (1961).
The infiltrometer simulates
rainfall, utilizing a single center-jet type nozzle held nine
feet above the ground by an aluminum tower.
The 7 LA nozzle
used in the present study delivered simulated rainfall at a
rate of approximately 10.5 cm. per hr. (4.2 in. per hr.) with
an energy of approximately 600 tons per acre in. at the soil
surface.
The nozzle produces a full-cone spray pattern and
wets an area 13 feet in diameter.
The plot frames surround
an area 3.81 x 3.81 feet or-1/3000 acre, have sides 6 in. deep,
and are driven into the soil about 2 in.
Accumulated runoff
is measured in a collection tank by an automatic water stage
recorder, and the rate of runoff is obtained from the slope
of the cumulative runoff curve.
The infiltration rate is
then computed as the difference between application rate and
runoff rate.
Infiltration measurements were made for each antecedent
condition for a period of 2% to 3 hours on the Ida and 1 to 3
hours on the Grundy to insure adequate measurement of
equilibrium rates.
In general, the nearly constant rate of
runoff was reached after 60 or 70 minutes.
40
Although considerable care was taken to obtain constant
rates of application, there were variations in spray intensity
of ± 0.25 cm. per hr. during a given measurement period and
up to ± 0.6 cm. per hr. on different runs.
The within-period
variation plus the error involved in taking the slope from the
cumulative runoff curve contributed to the variability en­
countered in plotting the infiltration rate curves.
Soil moisture sampling
Soil moisture determinations were made gravimetrically
on samples taken from the border of each plot immediately
prior to each infiltration measurement.
A composite of 2
samples was taken from both sides of the plot at 7.5 cm.
depth intervals to a depth of 30 cm. and at 15 cm. intervals
for the 30-150 cm. depth.
"Irrometer" tensiometers were used on the grass plots to
provide an approximate measure of the moisture content in the
surface 30 cm. of soil before infiltration measurements.
Three tensiometers were placed in the border outside the plot
frame at depths of about 5, 12 and 27 cm. (the depth given
corresponds to the center of the 4 cm. ceramic cup on the
tensiometer).
41
Soil sampling
For conductivity and bulk density measurements,
"undisturbed" soil cores, 7.4 cm. in diameter, were taken
with a power sampler (Buchele, 1961) from selected plots on
which infiltration measurements had been made.
Each 38 cm.
long core was encased in a series of 5 aluminum tube liners
when taken from the sampler, so that 5 sections of soil core,
each 7.6 cm. in length, could be easily cut apart with a
knife.
Cores taken from the 0-30 cm. depth for capillary con­
ductivity and moisture retention measurements were left in
the aluminum liners, treated with about 10 drops of formalde­
hyde to inhibit microbial activity, wrapped in a plastic bag,
and stored in ice cream cartons at 40 degrees F. until used.
Care was taken to identify the depth and vertical orientation
of each core.
On the Ida soil, duplicate core samples for bulk density
measurements were taken to a depth of 38 cm. on 6 plots and
to a depth of 76 cm. on 3 plots.
On the Grundy, triplicate
cores were taken on three plots to a depth of 38 cm., and
duplicate cores were taken in a pit to a depth of 150 cm.
The gravimetric moisture content of the Ida soil at the time
of core sampling was between 15 and 30 percent moisture
(depending on the plot and depth) and nearly constant at
42
about 40 percent on the Grundy soil.
The variation in mois­
tures on the Ida should not introduce error into the bulk
density data since the Ida soil shrinks and swells very
little.
Soil samples for organic matter and particle size dis­
tribution analyses were taken to a depth of 150 cm. on both
soils.
Four cores were composited on each of two plots to
give two samples for each depth.
The soil used for the aggregate stability and surface
sealing analyses was taken from the surface 10 cm. of freshly
tilled plots which had been wetted and allowed to reach
"field capacity" prior to sampling.
The soil was put through
an 8 mm. sieve and stored in a damp cellar where it was al­
lowed to dry slowly at about 65 degrees F.
Laboratory Procedure
Capillary conductivity and moisture retention
Capillary conductivity and moisture retention were de­
termined by the outflow technique of Gardner (1956) as modi­
fied by Kunze and Kirkham (1962).
The field cores used for these measurements were all
taken from one bromegrass plot on which infiltration measure­
ments had been made at three antecedent moisture levels.
necessity of obtaining duplicate measurements for a given
The
u
43
depth and the limitation in laboratory space required that
conductivity measurements be made on cores from one plot
only.
Since it was expected that the moisture flow charac­
teristics of the surface 30 cm. of soil would dominate the
infiltration process, the cores for conductivity measurements
were taken from this zone.
Eight cores, 7.5 cm. long, which
had been cut from two 30 cm. cores, were trimmed to a length
3
of 7.0 cm., giving a soil volume of approximately 300 cm. .
Individual pressure plate units were used for each core
sample.
Four commercial units, called volumetric pressure
plate extractors and made by the Moisture Equipment Company,
were used for one 0-30 cm. set of cores, and four similar
units constructed by the A.R.S., U.S.D.A. shop in Beltsville",
Maryland, were used for the other set of cores.
Although the
ceramic plates in both types of units had an air entry value
in excess of 2 atmospheres pressure, the water flow impedance
on the Moisture Equipment units was much greater than on the
Beltsville units.
Soil cores were placed on previously soaked ceramic
plates and allowed to saturate from below for 4 or 5 days.
To inhibit microbial activity about 10 drops of formaldehyde
solution (36.9 percent formaldehyde, 12.5 percent methanol)
was applied to each core and a few drops added to the buret
which received the outflow.
Although formaldehyde additions
to water could affect the surface tension of the soil water,
44
the effect was considered small.^
Other chemicals such as
toulene, mercuric chloride, and nitrophenol have been used by
other workers, but all have some disadvantages.
The formalde­
hyde appeared to inhibit effectively fungal growth on the soil
and in the water reservoir (buret).
Formaldehyde was added
to the buret every time the buret water was changed between
pressure steps.
After saturation, the units were sealed, and 25 cm.
water pressure was applied.
Outflow was not measured for the
first pressure step since plate impedance to flow was known
to invalidate the application of diffusion theory to the out­
flow process at pressures less than 25 cm. water.
All pres­
sures (i.e. suctions, in terms of negative soil-water
potential) are expressed as an equivalent height of a water
column since this unit was convenient to use in both the
conductivity and infiltration calculations.
Outflow measurements were made in a constant temperature
A 1 percent formaldehyde-plus-methanol solution could
be expected to have a surface tension of about 60 dynes per
cm., or about 83 percent that of pure water at 25 degrees C.
(Hodgman, 1960). If the soil contained 50 percent water at
saturation, the total volume of water in the core would be
150 cm.3. Since the 10 drops of formaldehyde solution added
is equivalent to about 0.5 cm.3, the concentration of formal­
dehyde-plus-methanol in the resulting soil solution would be
(0.5)(0.494)/150 or about 0.16 percent. After pressure is
applied and the soil becomes unsaturated, the formaldehyde
volatilizes and reduces the concentration further, so that
the effect of the added formaldehyde solution on surface
tension should be negligible.
45
room at 78 degrees F.
Small commercial heating units at­
tached to the top of the pressure units produced a temperature
gradient which reduced water condensation on the sides of the
unit.
Air diffusion through the plates was sufficient even
at low pressures to necessitate periodic purging of air bub­
bles from under the plate.
A pump developed by Kunze and
Kirkham (1962) was used to force water from the buret through
the hose and under the ceramic plate, forcing the accumulated
bubbles out through the measurement pipette.
It was neces­
sary to enlarge the grooves in the brass base plate beneath
the ceramic plate on the Moisture Equipment units to facili­
tate purging.
Pressure increments were chosen to cause between 2 and
10 ml. of outflow when possible.
It is necessary to keep out­
flow volumes small since the diffusivity is moisture depend­
ent, and the method assumes the constancy of D for small
pressure steps.
With the Kunze and Kirkham (1962) technique, the most
useful portion of the experimental outflow data is obtained
during the first few hours of outflow.
However, a period of
7 to 10 days usually elapsed before outflow ceased.
This
"late" outflow is not predicted by theory and generally re­
sulted in a poor fit of the experimental data to the theory
curves at large values of t.
The poor fit may be due to a
change in diffusivity as the moisture content of the soil
46
changes during outflow.
However, this effect does not appear
to be serious if the early portion of the data fits well.
Owing to the plate impedance effect mentioned previously,
conductivity values obtained with the ceramic plates in­
creased as moisture contents decreased until a pressure of
100 or 150 cm. water was reached, and thereafter the K values
decreased as moisture content decreased.
This apparent
double-valued K(0) relation was not in keeping with theory,
so another procedure was sought for the low pressure (high
moisture content) range.
Millipore filters which have a very uniform pore size
and little impedance to water flow were used successfully in
conductivity measurements by Kunze and Kirkham.
The Moisture
Equipment pressure units were modified to use nylon reinforced
millipore filters with 3 micron pore diameter in the place of
ceramic plates.
Perforated stainless steel plates seated in
plexiglass rings were constructed to support the millipore
filters.
Outflow measurements using the millipore filters were
accomplished on four soil cores taken from the same plot as
the cores used on ceramic plates.
Since it seemed advisable
to obtain more information on the surface 7.5 cm. of soil
than on greater depths, two cores from the 0-7.5 cm. depth
were used and one each from the 7.5 to 15 cm. and 15 to 22.5
cm. depths.
Outflow was measured for pressures from 0 to 130
47
cm. water.
Although some error could be expected at the zero
initial pressure, the measurement was made as accurately as
possible by applying 3 cm. water pressure to the units while
an equivalent back pressure was maintained in the outflow
pipettes.
Pressure sequences used for both ceramic and
millipore membranes are given in Tables 9 and 10, Appendix B.
Since infiltration is an absorption phenomenon, measured
values of K(0) and i|>(0) for the wetting process should be
more appropriate than desorption data for infiltration
calculations.
Although the outflow (desorption) procedure
developed by Gardner was not intended for use with absorption
measurements, an attempt was made to apply the same techniques
for absorption.
A 1 ml. pipette was adapted with two 4 cm.
pieces of glass tubing attached perpendicular to the pipette,
one near each end.
The tubing near the zero mark on the
pipette allowed introduction of a bubble into the pipette for
precise measurement of water volumes absorbed.
The second
tube near the outlet end of the pipette provided a bubble
trap and allowed atmospheric pressure in the pipette when a
bubble was introduced.
The water absorbed (volume vs. time)
after a small drop in pressure was measured in the pipette
first and later in a constant-head buret which served as the
water source.
Absorption measurements were made with both ceramic
plates and millipore filters on the same cores on which
48
desorption determinations had been accomplished.
After the
pressure had been raised to 500 cm. water on the 8 units
having ceramic plates, the pressure was then decreased step­
wise to 25 cm. water on the 4 Beltsville units for absorption
measurements, while desorption was continued on the 4 Moisture
Equipment units until a pressure of about 2000 cm. water was
reached.
It was thought that valid absorption measurements
might not be possible at pressures greater than 500 cm. water
owing to the discontinuity of water films, hence the above
procedure was used to provide low pressure absorption data on
one set of pressure units and high pressure desorption data
on the other set.
The experimental inflow data with ceramic
plates fit the theory curves very well suggesting that the
desorption theory was appropriate to the absorption phenome­
non.
Absorption measurements with millipore filters were
made following desorption, starting at a pressure of 130 cm.
water.
Although these inflow measurements at low pressures
provided a valid measurement of the t|>(0) relation for high
moisture contents, in general the data fit outflow theory
curves poorly, giving little information on K(9).
When the desorption-absorption process was completed,
soil cores were dried for 48 hours at 105 degrees C., and
moisture contents corresponding to the final pressure were
calculated on a volume basis.
The total outflow or inflow
for each pressure step was then used to calculate the moisture
49
content corresponding to a given pressure.
Particle size distribution
Particle size distribution was determined by the pipette
method (Kilmer and Alexander, 1949) using a mixture of Calgon
and sodium carbonate as a dispersing agent.
The following
size groups were separated: less than .002 mm. (clay), .002
to .020 mm. (fine silt), .020 to .050 mm. (coarse silt), and
.050 to 2.00 mm. (sand).
Organic carbon
Total carbon was determined on both soils by the dry
combustion method, while inorganic carbon was determined on
the Ida soil only, owing to its high calcium carbonate con­
tent, by measuring CO^ evolution from the soil after the
addition of weak acid.
(1957).
Both methods are described by Black
The organic carbon content on the Grundy was as­
sumed to equal total carbon, while on the Ida organic carbon
was given by the difference between total and inorganic
carbon.
Aggregate stability
Aggregate stability was measured by a wet sieve analysis.
Air dry soil was divided into the following aggregate size
groups by dry sieving: greater than 2.83 mm., 2.83 to 2.00
50
mm., and 2.00 to 1.00 mm.
A 25 g. sample of aggregates from
each size group was wetted for about 3 minutes under vacuum
(60 cm. mercury) and then wet sieved for 5 minutes in a
mechanical sieving apparatus which moves a nest of sieves
under water in a 2.5 cm. vertical stroke at the rate of 30
oscillations per minute.
The relative stability of different
sized aggregates was measured by determining the size frac­
tions into which each dry-aggregate group broke down upon •
wet sieving.
Surface sealing
When cultivated soils are subjected to rainfall impact
two phenomena are mainly responsible for the decrease in
infiltration rate with time:
(a) the rapid decrease in
potential gradient near the surface and (b) the decreased
conductivity of the immediate soil surface due to aggregate
breakdown and surface sealing.
Although a measure of surface
sealing was obtained with field infiltration measurements, an
experiment with controlled conditions was needed to evaluate
the effect of initial moisture content on the breakdown of
aggregates and subsequent sealing of the surface due to rain­
fall impact.
The laboratory rainfall simulator used in this
experiment (Mutchler and Moldenhauer, 1962) was especially
well adapted to a study of surface sealing since the usual
decrease in potential gradient with time was prevented by
51
maintaining a constant suction at the lower boundary of a
shallow aggregate layer.
Thus, after the initial wetting of
the aggregates by "rainfall", the measured decrease in intake
rate with time could be attributed to a decrease in the con­
ductivity of the surface as sealing occurred.
The rate at
which the intake rate decreased with time and the erosion
rate with time were.used as indices of aggregate breakdown
and surface sealing for the two soils at three antecedent
moisture levels.
The rainfall simulator consisted of a rotating water
reservoir in which a constant water head was maintained over
a system of capillary tube droppers.
The moving droppers
produce "rainfall" at a rate of approximately 9.0 cm. per hr.
with a well distributed pattern at the soil surface.
A soil
pan, 11.5 in. wide, 17.5 in. long and 5 in. deep holds the
soil and provides a means of measuring runoff and erosion
rates by volume collections at 5 minute intervals.
The soil
pan contains a layer of micro-glass beads which when saturated
and then subjected to a suction at the bottom of the pan will
maintain a suction up to about 250 cm. water before air entry
occurs.
Soil samples which had been taken from the field (see
section on soil sampling) and allowed to dry in a moist cellar
were sieved with a 2 mm. sieve to obtain aggregates greater
than 2 mm. in diameter.
The moisture content of the aggre­
52
gates at the time of sieving was used for the intermediate
initial moisture level in the simulator study.
Drier aggre­
gates were obtained by further air drying in the laboratory.
A wet initial condition was achieved by wetting the aggregates
in the soil pan under vacuum for about one hour and then ap­
plying a suction of 200 cm. water to the bottom of the soil
for about 4 hours.
A preliminary trial with tensiometers
showed that after application of suction to the bottom of a
saturated 7.5 cm. layer of beads, a suction of 200 cm. water
was reached near the surface of the beads in about 3 minutes.
A shallow layer of aggregates, 15 to 18 mm. deep, was
placed over the beads with care being taken to prevent segre­
gation of large and small fractions.
The surface was leveled
and the aggregates tamped lightly to produce a relatively
stable layer.
A dry, double layer of gauze between the soil
and beads prevented rapid movement of water from the wetted
beads (at 200 cm. water suction) into initially dry soil.
Duplicate measurements were made on each soil and mois­
ture condition combination.
On two runs with the Grundy
aggregates one duplicate (Pan B in both cases) was lost owing
to an air leak in the beads.
Although data for only one pan
were obtained on these two runs, the precision of other
duplicate measurements indicated that the data from a single
pan should provide a reliable estimate of infiltration and
erosion for the conditions being studied.
53
Calculation of Infiltration from Conductivity Data
When a partial differential equation is used to express
a physical process with a given set of initial and boundary
conditions, numerical methods are generally used to obtain a
solution.
The most commonly used method is called the "method
of finite differences" (Sokolnikoff and Redheffer, 1958).
The
differential equation is replaced by an approximating dif­
ference equation, and a set of discrete points defines the
continuous region in which the solution is desired.
The
problem is thus reduced to the solution of a system of alge­
braic equations in many unknowns.
Iterative techniques have
been devised to solve such systems, and high speed computers
can be used to handle the laborious calculations.
In the present study the equation to be solved is Equa­
tion (11) subject to the initial and boundary conditions,
t = 0, x > 0, 9 = f(x)
x = 0 , t > 0, e=©
0
,
where 0 = f(x) represents the measured antecedent moisture
profile.
It is assumed that the value of K at the boundary
between the two soil horizons is the mean K for the two hori­
zons and that i|> is continuous across the boundary.
The method
of finite differences was applied to the moisture diffusion
54
equation by Hanks and Bowers (1962a) who developed a computer
program to accomplish the necessary calculations.
Some of
the pertinent steps are shown in the following discussion.
The finite difference approximation of
39
at
=
j_rKif
axLKax-
which is Equation (7) in the literature review, is as follows:
Vi'9lvi-l
At
= K
^i-l,i
+
^i-1,j-1 *
2G
" ^i,j " ^i,,i-l)
(18)
K
(^ivi
+
^1,1-1
+ 26
" ^i+l,i " ^i-hlvi-l)
2(AX)2
.
where i|) is the suction, K is the conductivity, G is the
gravitational term, At is the time increment, and the sub­
scripts "i" and "j" refer to distance and time, respectively.
G = ax for vertical infiltration.
By knowing the boundary
conditions and values of K, an equation can be written for
each depth increment, involving the unknowns 9^ j and
(i = 1, 2, 3, ... n-1).
j
The initial conditions at the be­
ginning of each new time increment supply values of
55
and i|)j_ j_]_ for all depth increments.
A series of "n" equa­
tions having more than "n" unknowns is then formed but addi­
tional information is needed to obtain a solution.
If the
i|>(©) relation is known and can be assumed to be unique, then
the left-hand-side of Equation (18) is approximated by
• •
^
—
Ay • •
-±*=±C. , ,
Ui jJ4
At
?
where C. • i/ = (4r)
i,J-4
d*
(19)
and is taken from the measured o|>(0)
curve for an estimated value of 0.
Substitution of (19) in
(18) yields the series of "n" equations with
11 n"
unknowns for
which solutions can be obtained.
Twenty depth increments of constant length, ax - 2cm. were
used for all cases studied.
The calculated wetting front
did not exceed 30 cm. deep in any of the problems calculated.
The time increment, At? was varied automatically in the com­
puter program to allow for a constant amount of water entry
at the surface for each time increment.
The conductivity, K, was estimated as follows:
îî:î;]:î where
(20)
56
Da6
Although in the present study, K values had been obtained
from outflow measurements by the equation, K = D[Q/VAp] where
Q is the total outflow from a soil core of volume V for a
pressure step of AP> experimental values of D were used in
the computer program, and K was calculated by the computer
using Equations (20) and (21).
The value of
j f r o m
Equation (21) is an "average" value which Hanks and Bowers
(1962a) found to give better results than where a similar
expression in K was used.
The solution of a given problem involved the stepwise
solution of twenty algebraic equations, starting with the
equation corresponding to the first distance increment and
progressing to the last increment, each equation being
evaluated for the first time increment, j = 0 to j = 1.
The process was then repeated for each time increment to
provide infiltration data corresponding to a cumulative time
of 70 minutes.
Each problem required about one hour for
computation on the IBM 1620 computer.
The values of suction, moisture content, cumulative in­
filtration, infiltration rate, length of time increments,
57
and cumulative time were printed out for each depth increment
at frequent time intervals.
The cumulative infiltration, CI,
and infiltration rate, I, were computed as follows:
n
n
( C I ) . - 2 9 -1 - A * - S 9 1
. Ax
J
i=i 'J
"l—i '0
i=l
T
-
"
(^0,1
+
*0vi-l + 2G
"
(22)
" 4j,i-l)K%,i
(23)
2AX
Note that the expression for CI gives the accumulated mois­
ture in the profile at a given time as did the expression
xd9 in Philip's (1957b) analysis.
It is interesting to
Sn
note that the expression for infiltration rate, Ij, in (23)
reduces to I. = K(9 ,) when the first increment of soil is
j
sax
saturated.
Since the intake rate corresponding to K
^
should be the minimum rate reached if there are no restricting
zones in the profile, unsaturation at depths greater than 1
cm. would necessarily exist as long as the condition I > K
^
persisted.
On the basis of laboratory results the "soil profile"
used in computations was considered to consist of 2 horizons,
the 0-15 cm. surface soil and the 15-40 cm. subsoil, the 2
58
horizons having a distinct boundary.
The D and ij> data used
for the surface 15 cm. were obtained on field cores from the
same depth, while data for the 15-40 cm. depth were measured
on 15-22.5 cm. field cores.
Infiltration computations were made for 7 different
problems with the following conditions:
(a) Dry antecedent moisture , 2 horizons, desorption
data
(b)
Moist antecedent moisture, 2 horizons, desorption
data
(c)
Wet antecedent moisture, 2 horizons, desorption
data
(d) Dry antecedent moisture, 2 horizons, absorption
data
(e)
Moist antecedent moisture, 2 horizons absorption
data
(f)
Uniformly wet (35 percent), 2 horizons, desorption
data
(g)
Wet antecedent moisture, 1 horizon (using 0-15 cm.
data for the entire 0-40 cm. depth), desorption
data
The 7 conditions were to provide the following information:
1.
The dry, moist, and wet conditions in (a), (b), (c)
(d), and (e) correspond to the field moisture condi
tions for which infiltration rates were measured.
These computed rates are compared with field rates
to evaluate the predictive value of the laboratory
and computational procedures.
59
2.
A comparison of (a) and (d) with the field measure­
ment for the dry condition, and a comparison of (b)
and (e) with the field measurement for the moist
condition provide an evaluation of the merits of
desorption and absorption data obtained in the
laboratory for use in theory calculations.
3.
A comparison of (f) and (c) is used to estimate the
effect of the higher moisture content in (f) at
depths below 4 cm. on the infiltration rate and
cumulative infiltration.
4.
Condition (g) gives a measure of the error in predic­
tion introduced by considering the entire profile to
have the K and i|) properties of the 0-15 cm. depth.
60
RESULTS AND DISCUSSION
Physical Properties of Soil Profiles
The Ida and Grundy soils, although both loess derived,
differ strikingly in profile development as is shown by the
bulk density, organic carbon and particle size distribution
data in Tables 1 and 2.
The Ida silt loam, classified as a Regosol, was known to
have a relatively uniform texture with depth, and was chosen
for the present study for this reason.
Data in Table 1 show
that organic carbon and clay contents are highest in the sur­
face 22.5 cm. of soil and drop to relatively constant values
for greater depths.
The bulk density of the surface 7.5 cm.
is only slightly greater than at greater depths in the
profile.
The Grundy silty clay loam (Table 2) has a distinct Bhorizon with higher clay content and lower organic carbon
content than in the A-horizon.
In contrast to the Ida soil,
the bulk density of the Grundy is lowest in the surface soil
and increases gradually with depth to about 1.55 g./cm.
75 cm.
3.
at
61
Table 1.
Bulk density, organic carbon content, and particle
size distribution of Ida silt loam with depth
Soil
depth
Bulk
density
cm.
g./cm.^
0
-
7.5
7.5- 15
Organic
carbon
Particle size distribution
Clay F. silt C. silt Sand
%
%
%
%
%
1.24
1.69
21.6
27.8
46.5
4.1
1.21
1.63
26.6
47.2
5.0
1.21
21.2
21.9
15.6
27.4
45.4
5.3
46.0
5.5
46.1
5.7
22.5- 3 0
1.19
1.46
0.75
30
- 45
1.17
0.33
14.6
32.9
33.6
45
- 60
1.19
0.22
14.6
29.4
50.7
5.3
60
- 75
1.21
-
12.9
32.6
50.3
4.2
75
- 90
-
-
15.4
90
-105
-
-
17.6
30.8
30.6
49.0
48.1
4.8
3.7
105
-120
-
-
16.3
31.0
43.1
9.6
120
-135
-
-
13.8
30.8
47.5
7.9
135
-150
-
-
14.3
29.2
46.8
9.7
15
- 22.5
Aggregate Stability and Surface Sealing
Aggregate stability
Although two soils may have a similar size distribution
of aggregates after tillage, aggregates of the same size on
different soils or of different sizes on a given soil may
exhibit varying degrees of water stability.
Data in Table 3
show that the Ida and Grundy soils had a similar proportion
62
Table 2.
Bulk density, organic carbon content and particle
size distribution of Grundy silty clay loam with
depth
Soil
depth
Bulk
density
cm
g./cm.^
%
%
%
%
%
1.28
2.39
30.5
28.9
37.1
3.5
1.34
2.38
29.4
33.5
36.2
5.4
1.32
29.0
30.6
32.7
34.3
32.5
30.1
3.2
3.1
34.9
32.7
24.2
19.3
2.7
3.5
0
•
-
7.5
7.5- 15
Particle size distribution
Clay F. silt C. silt Sand
2 2 . 5- 3 0
1.34
2.02
1.73
30
- 45
1.36
1.29
45
- 60
1.39
0.98
38.2
44.5
60
- 75
1.49
0.50
39.9
38.9
18.2
3.0
75
- 90
1.54
38.0
39.8
19.8
2.4
90
105
-105
1.58
35.6
40.0
21.9
2.5
-120
1.55
34.8
40.3
2.6
120
-135
1.55
0.18
31.6
2.7
135
-150
1.54
0.10
32.5
40.2
44.0
22.3
25.5
21.9
1.6
15
- 22.5
Organic
carbon
-
0.37
-
of aggregates in the >2.8 mm. size group prior to wet sieving.
However, the proportion of these large aggregates which re­
sisted breakdown when subjected to wet sieving was 59.3 per­
cent for the Ida and 75.0 percent for the Grundy.
Also, of
the 25 percent of the Grundy aggregates which passed through
the 2.8 mm. sieve with wet sieving, less than half broke down
to the <1.0 mm. size.
A much larger proportion of the Ida
aggregates was dispersed, indicating the lower water stability
Table 3.
Size distribution and relative water stability of Ida silt loam and
Grundy silty clay loam aggregates
Time of
measurement
Before wet
sieving
Soil
\3
Percent aggregates in each size group (diameter in mm.)
>2.8
2.8-2.0
2.0-1.0
Ida
31.4
12.0
19.8
Grundy
32.6
19.8
29.3
>2.8
After wet
sieving
aThe
2.8 - 2 . 0
2.0-1.0
2.8-2.0
2.0-1.0
2.0-1.0
Ida
59.3
8.6
4.0
29.4
27.4
31.2
Grundy
75.0
8.7
5.1
66.1
18.6
66.5
dry soil was first sieved to obtain the percentage aggregates in each
size group; then each of the dry aggregate size groups was wet sieved separately,
giving the data in the lower half of the table. The percent in the <1.0 mm. size
group is not shown in the table, but may be obtained by difference.
64
of the Ida soil.
The greater stability of the Grundy aggre­
gates is also evident in the "after wet sieving" data for the
2.8-2.0 mm. and 2.0-1.0 mm. size groups.
The wet sieving data
suggest that the >2.8 mm. aggregates are more stable than the
two smaller aggregate sizes on both soils.
However, the >2.8
mm. group included aggregates of sizes up to 8 mm., a much
larger size range than in either of the smaller size groups.
Hence aggregates >2.8 mm. resulting from the breakdown of a
large aggregate could still remain on the >2.8 sieve.
It is
apparent that aggregates in the 2.8-2.0 mm. and 2.0-1.0 mm.
size groups have about the same water stability on a given
soil.
Although Grundy aggregates were shown to be more water
stable than Ida aggregates in the wet sieving analysis, such
a measure of stability does not necessarily imply a resistance
to aggregate breakdown and surface sealing under rainfall
impact.
Surface sealing
A measure of the effect of initial moisture content on
aggregate breakdown and surface sealing using the laboratory
rainfall simulator is given by the infiltration and erosion
rate curves in Figures 2 and 3.
Curves in the upper half of
Figure 2 show that dry Ida aggregates maintained a higher in­
filtration rate for times up to 50 minutes than did either
65
the moist or wet aggregates.
The Grundy aggregates, on the
other hand, appeared to be most stable in the wet condition,
i.e. after being initially wetted and then held at a low
suction prior to water application (lower half of Figure 2).
The difference between the "dry" and "moist" curves for the
Grundy is slight if the curves are translated along the
horizontal axis until they coincide.
The shape of the infil­
tration curve is considered a more important index of surface
sealing than is the time that runoff began (i.e. when infil­
tration was first measured), since the latter reflects both
the time needed to wet the aggregates and the rate of sealing.
The equilibrium infiltration rates, shown by the data points
on the 50 minute ordinate, were about 5.0 cm. per hr. for the
Ida and 4.0 cm. per hr. for the Grundy.
Since clay particles
might be expected to form a more dense seal than silt parti­
cles, it is not surprising that the seal formed on the Grundy
was less permeable than on the Ida.
The higher erosion of the Grundy soil at all three
moisture conditions as compared to similar conditions on the
Ida reflects the lower infiltration rates of the Grundy
(Figures 2 and 3).
An effect of initial moisture content on
erosion rate is evident on the Grundy but not on the Ida.
Erosion rates remained high throughout the time of measure­
ment on both soils even though a surface seal of fairly con­
stant permeability had been formed.
Apparently the surface
Figure 2.
Effect of initial volumetric moisture content
of Ida and Grundy soil aggregates on surface
sealing as indicated by rate of water infil­
tration (determined in the laboratory using a
rainfall simulator)
67
10
-= = = = V
8
® IDA
:
6
:o~
tr.
=0:
4 - • DRY (2%)
o
2
-
A WET (19%)
UJ
5
q:
o MOIST (9%)
0
30
20
10
40
50
10 -f
(D GRUNDY
o
f= 8
<
cr
A
•"
h 6
-A.
^ 4
•o.
• DRY (4%)
-A.
~A"<5"
o MOIST (14%)
2
a WET (25%)
0
0
10
20
30
TIME (MINUTES)
40
50
Figure 3.
Effect of initial volumetric moisture content
of Ida and Grundy soil aggregates on surface
sealing as indicated by soil erosion rate
(determined in the laboratory using a rainfall
simulator)
69
5 4 -
• DRY (2%)
® IDA
o MOIST (9%)
A
WET (19%)
5 3
o
O
X
LU
2
-
ÛT
O
<
êf
X
t
CO 0
o
1
20
5 -
o 4 -
QL
1
40
• DRY (4%)
1
r
60
80
(D GRUNDY
o MOIST (14%)
if)
txJ
,o>
A
WET (25%)
'
.-•-•x
/\
v
3 -
_J
O
CO
2
-
0
i
40
1
60
TIME (MINUTES)
1
r
80
70
seal was not a stable layer but was constantly eroding away
and being reformed by the continual breakdown of aggregates.
Below the 1 to 3 mm. dense layer formed by "rainfall" impact
the soil aggregates maintained their original structure.
Thus, surface sealing continues to be a dynamic process even
after an equilibrium infiltration rate is reached.
The opposite effect of initial moisture on the sealing
of the Ida and Grundy aggregates might be explained by a dif­
ference in the materials forming the bonds between particles.
The Ida aggregates, having a low clay content but a high
calcium carbonate content, may be stabilized with drying by
the cementing effect of calcium carbonate.
Such cementing
would not occur on moist or wet aggregates.
The Grundy
aggregates, on the other hand, contained no calcium carbonate,
but were relatively high in clay--.content.
Thus, for the
Grundy, the lower stability of the dry aggregates may be as­
sociated with air entrappment upon wetting, the drier aggre­
gates being readily broken down by this means.
Although
Cernuda et al. (1954) concluded that aggregates were most
subject to slaking in water when initially dry, none of the
soils which they studied contained large amounts of calcium
carbonate.
These comments on the mechanisms of bonding and
breakdown of aggregates are obviously speculative, and neces­
sarily so, since little is known about the nature of soil
aggregate bonds.
71
A measure of surface sealing was also obtained in the
field infiltration measurements.
These data will be dis­
cussed later when field results are presented.
Moisture Retention, Diffusivity,
and Capillary Conductivity
Moisture retention
The suetion-moisture content relationships for the sur­
face 30 cm. of the Ida soil are shown in Figure 4.
Hysteresis
in the desorption-absorption cycle is evident at all four
depths, though more pronounced in the surface 0-15 cm.
The position of the curves in the four graphs with re­
spect to the ordinate scale shows that the moisture content
at a given suction was higher in the surface 15 cm. of the
profile than at greater depths.
The uniformity of the 0-15
cm. depth with respect to moisture retention is apparent in
Figures la and lb.
The uniform clay and organic carbon contents with depth,
below 22.5 cm. (Table 1), indicate that moisture retention
and capillary conductivity data for the 22.5-30 cm. cores
should provide a good estimate of retention and conductivity
at greater depths.
The x|j(9) curves in Figure 5 are "average" curves from
which data were taken for the computer calculations of
Figure 4,
Moisture content versus suction, t|)(9), for
desorption and absorption on field cores of
Ida silt loam at four depths
73
. DESORPTION
o ABSORPTION
0-7.5 CM. DEPTH
7.5-15 CM. DEPTH
- 50
o
>
> 40 CD
UJ
U
30 -
OC
UJ
Q_
UJ
H
Z
O
20
200 400
50 -
0
200 400
©
o
©
\
UJ
cr 40 -
3
hen
30 15-22.5 CM. DEPTH
20
0
200 400
22.5-30
0
200 400
SUCTION (CM. WATER)
Figure 5.
"Average" ^(0) curves used in theory calculations of infil­
tration; laboratory data from the 15-22.5 cm. depth were
used for the >15 cm. depth in computations
1
o
> 50
I
I I I I I
DESORPTION
>
m
ABSORPTION
40 h-
z
DESORPTION
UJ
o 30
ABSORPTION
-4
Ui
o
UJ
a:
3
Ic
—
0 - 1 5 CM. DEPTH
20
o
> 15 CM. DEPTH
10
T
0
1
1|I I I |
I
10
1
1
1—| I I I|
I
1
100
SUCTION (CM. WATER)
1
1—| I I I |
I
1000
76
infiltration.
These curves represent moisture retention data
obtained on both the ceramic plates and millipore filters.
Individual values are given in Tables 9 to 14, Appendix B.
The 0-15 cm. depth curves in Figure 5 were obtained from the
combined data for the 0-7.5 cm. and 7.5-15 cm. depths, while
data from the 15-22.5 cm. depth were used to represent the
remainder of the profile.
Diffusivity and capillary conductivity
The relationship of water diffusivity and moisture con­
tent for the Ida soil is shown in Figure 6, and the corre­
sponding capillary conductivity curves in Figure 7.
curves were fitted visually to the data points.
The
Although
unique curves could be established for the 15-22.5 and 22.5-30
cm. depths, the 0-7.5 and 7.5-15 cm. depth data appear to fit
approximately the same curve.
There is a tendency for the
0-7.5 cm. data to fall to the right of the 7.5-15 cm. points
at moisture contents of 35 to 50 percent, especially in
Figure 7.
The conductivity data are less divergent than the
diffusivity data for the 0-15 cm. depth.
Three items concerning the data in Figures 6 and 7 are
noteworthy:
(a)
The shape of the curves is consistent with published
data. At lower moisture contents the curves are
very nearly exponential (linear on semi-log plot),
but tend toward an inverted S-shape at high moisture
77
contents. Gardner and Mayhugh (1958) found that
an assumed exponential relation between diffusivity
and moisture content gave good results on the soils
they studied. In another paper Gardner (1959)
shows a plot of D vs. 9 on a semi-log scale, and
although he had drawn a straight line through the
data points, the curves appear to be of a similar
shape as those obtained in Figure 6. A calculated
D vs. 0 curve shown by Philip (1955) is also "of
this shape.
(b)
Desorption and absorption data fall roughly on the
same curves, especially for the two greater depths.
This suggests a unique K(0) relation even though
hysteresis exists in the i|>(0) curves of Figure 4.
However, for the surface 0-15 cm., the absorption
data in both Figures 6 and 7 tend to lie slightly
to the left of desorption points. It should be
noted that since D = K(3i|)/30), if the K(9) relation
is unique, then the D(0) relation cannot be unique,
owing to hysteresis in ij>(0). The divergence of
D(0) from uniqueness is then dependent on the
magnitude of the difference in the x|)(©) curves for
desorption and absorption. The wide spread of D
values for the 0-15 cm. depth in Figure 6 as com­
pared to the corresponding K values in Figure 7
supports the above reasoning. Also, the relatively
good fit of the D(0) curves for the two greater
depths to both desorption and absorption data
should be expected since hysteresis in a|>(0) is less
for those depths. Data obtained by Nielsen and
Biggar (1961; indicate a unique K(0) relation in
disturbed samples after initial soil consolidation.
Also, Gardner (1959) obtained D values for absorp­
tion which were 2 to 3 times as great as those for
desorption at a given moisture content, i.e. the D
curves for absorption fell to the left of those for
desorption. Hence, the data in Figures 6 and 7
appear to be consistent with published results.
(c)
The D and K values obtained with millipore filters
correspond well with values obtained with ceramic
plates. The degree of correspondence is surprising
considering that different field cores were used to
obtain the two sets of data.
D values taken from the 0-15 cm. and 15-22.5 cm. curves
in Figure 6 were used in the computer calculations of
Figure 6.
Water diffusivity versus moisture content,
D(9), for desorption and absorption on field
cores of Ida silt loam at four depths; data
points corresponding to the symbols in
parentheses in the legend were obtained with
millipore filters, while all others were
measured on ceramic plates
79
10 4 _
DEPTH
© (O)
®
®)
© (•)
10 3
a:
x 102
S
o
t 10'
to
3
Li.
U.
O
ÛC
22.5 - 30-
UJ
h-
I
10° — CM. DEPTH
0-15 CM.
10"1
DEPTH
—
15-22.5
CM. DEPTH
10-2
0
10
20
30
40
50
MOISTURE CONTENT, (% BY VOL.)
60
Figure 7.
Capillary conductivity versus moisture content,
K(Q), for desorption and absorption on field
cores of Ida silt loam at four depths; data
points corresponding to the symbols in paren­
thesis in the legend were obtained with
millipore filters, while all others were
measured on ceramic plates
81
0 _
DEPTH
DES.
® (O)
ABS.
3 (®)
HT' —
>-2_
22.5 -30-
,-4
CM. DEPTH
15-22.5 —
0-15 CM. DEPTH
CM. DEPTH
,-7,
0
10
20
30
40
50
MOISTURE CONTENT, (% BY VOL.)
60
82
infiltration and are given in Table 15, Appendix B.
Field Infiltration
Antecedent moisture effect
The antecedent moisture profiles and corresponding in­
filtration curves for the grass sod on the Ida soil are shown
in Figures 8 and 9 (Replications I and II, respectively).^
The moisture content versus depth curves show the mois­
ture distribution in the profile for each of the designations,
"dry", "moist", and "wet".
These terms refer to the relative
moisture content of the surface soil at the beginning of the
infiltration measurement.
The infiltration rate curves in Replicate I (Figure 8)
are distinctly different for each antecedent moisture, with
the infiltration rate at a given time being inversely related
to antecedent moisture content.
On the second replicate
(Figure 9) the infiltration rate for the "dry" condition was
highest, while the rates for the "moist" and "wet" conditions
were nearly equivalent.
These apparent equivalent infiltra­
tion rates on Replicate II may be the result of having
"'"Data for the third replicate are not graphed since one
infiltration run was lost, but cumulative infiltration and
equilibrium infiltration rates for all replicates are shown
later in Table 4.
Figure 8.
Field measured infiltration rates of Ida silt loam (grass
cover) at three antecedent moisture levels, Replicate I;
data points on the 70 minute ordinate represent the average
rate after 70 minutes
MOISTURE %
(BY VOL.)
10 20 30 40
IDA I - l
30
S 10.0 -
o *0
o DRY
• MOIST
a WET
8.0
UJ
90
CL
UJ
120
O
< 6.0
w
-1
•
-/
a
V
150 I
V
tK
Z
2 4.0 I<
OC
H- 2.0
0.0
0
10
30
40
%
TIME (MINUTES)
50
60
70
Figure 9.
Field infiltration rates of Ida silt loam (grass cover) at
three antecedent moisture levels, Replicate II; data points
on the 70 minute ordinate represent the average rate after
70 minutes
MOISTURE %
(BY VOL.)
10 20 30 40
IDA H-l
x 10.0 H
— 8.0
6/
DRY
MOIST
WET
H
UJ
Â
uj 120
S 6.0 H
p 4.0 S
—l
u_
2.0 -
o.o -4
10
20
30
40
TIME (MINUTES)
50
60
7(
88
somewhat similar moisture profiles for the "moist" and "wet"
conditions.
A comparison of moisture profiles in Figures 8
and 9 shows that the "wet" profile in Figure 8 has a much
higher moisture content at depths greater than 60 cm. than
does the "wet" profile for Replicate II in Figure 9, possibly
explaining the difference in results on the two replicates.
The higher moisture content at greater depths in Replicate I
resulted from an earlier "wet" run in which a leak developed
around the plot frame, necessitating another "wet" run.
Despite the inconsistency in results on the two replicates,
the influence of antecedent moisture on infiltration rates
is evident.
•In Figures 10 and 11 infiltration curves for the tilled
surface soil, both screen covered and bare, provide a measure
of the magnitude of tillage and s.urface sealing effects for
comparison with antecedent moisture effects.
The "moist"
curves in Figures 8 and 9 are shown again in Figures 10 and
11 as the "grass" curves.
A consistent effect of tillage is to prolong the time
before runoff occurs (Figures 10 and 11).
In both replicates
the "screen" curve is above the "grass" and "bare" curves for
the entire 70 minute period, showing the capacity of the
tilled soil to receive water when the surface remains
permeable.
In addition, the location of the "bare" curve
relative to the "screen" curve is nearly the same on both
Figure 10.
Field infiltration rates of Ida silt loam as influenced by
tillage (screen versus grass) and surface sealing (bare
versus screen), Replicate I; data points on the 70 minute
ordinate represent the average rate after 70 minutes
IDA I
a SCREEN
° BARE
• GRASS
< 6.0
vo
O
0.0
J
20
30
40
TIME (MINUTES)
50
Figure 11.
Field infiltration rates of Ida silt loam
tillage (screen versus grass) and surface
versus screen), Replicate II; data points
ordinate represent the average rate after
as influenced by
sealing (bare
on the 70 minute
70 minutes
IDA H
a SCREEN
tr 10.0 X
\
5
o 8.0 -
o BARE
• GRASS
LU
\—
< 6.0 o:
z
o
h- 4.0.<
cc
1_l 2.0 -
I
O
z
0.0 C
10
20
30
40
TIME (MINUTES)
50
60
7(
93
replicates, the water intake rate of the tilled soil being
greatly reduced by rainfall impact on the bare surface.
The effects of tillage and surface sealing on infiltra­
tion rates are clearly evident in Figures 10 and 11.
A com­
parison of Figure 8 and Figure 10 indicate that the magnitude
of the antecedent moisture effect on infiltration may be as
great or greater than the effects of tillage or surface
sealing.
This is an important result in view of past and
present efforts to characterize soils and soil management
practices with regard to infiltration rate.
Although the
moisture effect was less and the tillage effect greater in
Replicate II than in Replicate I, the moisture effect is of
sufficient magnitude to necessitate evaluation before effects
of management practices can be accurately estimated.
Infiltration rates for one replicate of the Grundy soil
are given in Figure 12.
Data for the other replicate were
nearly the same as in Figure 12, except that the infiltration
rate for the
11
bare" condition decreased to the equilibrium
value more rapidly.
The antecedent moisture level for all
three surface conditions was near the "field capacity" or
roughly equivalent to the "wet" condition on the Ida soil.
The similarity of the infiltration rate curves with
"grass" cover for the Ida (Figure 11) and Grundy (Figure 12)
suggests that morphological differences between soils may not
influence infiltration rates for small times as much as
Figure 12.
Field infiltration rate of Grundy silty clay loam as in­
fluenced by tillage (screen versus grass) and surface sealing
(bare versus screen), Replicate I; data points on the 70
minute ordinate represent the average rate after 70 minutes
GRUNDY X
ûf 10.0
o 8.0
LU
6.0
O 4.0
a SCREEN
h- 2.0
LL
o BARE
• GRASS
10
20
TIME (MINUTES)
96
antecedent moisture and surface conditions on a given soil.
The screen covered surface on the Grundy (Figure 12)
maintained a very high infiltration rate throughout the mea­
surement period.
It was discovered after the run that water
was infiltrating through the permeable surface and then
moving laterally through the tilled soil and out around the
steel retaining plate at the front of the plot.
This problem
could exist on any soil in which the surface few inches are
much more permeable than the underlying soil.
Despite the
error, these data do indicate the high permeability of the
tilled surface which has been protected from rainfall impact.
The "bare" surface of the Grundy, although gradually
sealed by rainfall impact, maintained a higher infiltration
rate than did the grass sod for the 70 minute period shown.
Similar results were obtained on one replicate of Ida (Figure
11), but not on the other two replicates, in which the infil­
tration rate curve for the bare surface dropped rapidly below
the curve for the grass sod (e.g. Figure 10).
The variability
of the results on the Ida soil emphasizes the difficulties
encountered in characterizing soils with respect to infiltra­
tion.
Although one is generally interested in knowing the
shape of the infiltration rate curve for a given soil and
condition, equilibrium infiltration rates and cumulative in­
filtration values are also informative.
These data are given
97
for the Ida soil in Table 4 and for the Grundy soil in Table
5.
Mean equilibrium rates for the three antecedent moisture
levels in Table 4 (significant at the 0.25 level) indicate
that equilibrium rates achieved after 2.5 to 3 hours of mea­
surement were inversely related to the antecedent moisture
level.
Similar results were found by Free et al. (1940).
In
previous studies using simulated rainfall on a cultivated
Marshall silt loam it was found that equilibrium rates on dry
and subsequent wet runs were consistently equal (Green et al.,
1961).
Apparently, the equality of rates on the Marshall re­
sulted from surface sealing, the permeability of the seal
having reached a relative constant value by the end of the
dry run.
Thus, on the Marshall the wet run gave, essentially,
a measure of the permeability of the surface seal formed
during the dry run and did not reflect moisture potential
differences.
In the present study, the grass surface was
stable and differences in infiltration rate for the three
antecedent moisture levels can be assumed to be the result of
differences in potential gradients within the soil.
This
effect will be discussed in more detail when calculated mois­
ture profile data are presented.
The 60 minute cumulative infiltration values for the
three moisture levels (Table 4) are consistently higher for
low moisture contents, as would be expected.
The differences
98
Table 4.
Infiltration indices for Ida silt loam as influ­
enced by various antecedent conditions
Antecedent
condition
Grass sod
Dry
Moist
We+
Tilled, covered
Tilled, bare
aValues
Rep.
Equilibrium
infiltration
rate
cm./hr.
Cumulative
infiltration
in 60 min.
cm.
I
II
III
3.6
2.8
4.1
5.9
4.7
5.5
Mean
3.5
5.4
I
II
IIIa
2.8
2.0
(3.6)
5.0
3.9
(4.8)
Mean
2.4
4.4
I
II
III
1.5
2.0
3.3
3.8
3.5
4.2
Mean
2.3
3.8
I
II
III
2.5
3.6
1.9
6.2
7.4
5.5
Mean
2.7
6.4
I
II
III
1.3
1.0
1.0
Mean
1.1
3.4
4.8
4.2
4.1
in parentheses for Replicate III are calculated
"missing plot" values' used in statistical analysis. No mea­
sured values were obtained owing to a clogged nozzle during
the measurement period.
99
Table 5.
Infiltration indices for Grundy silty clay loam as
influenced by three surface conditions
Surface
condition
Rep.
Equilibrium
infiltration
rate
cm./hr.
cm.
o
Tilled, bare
00
Tilled, covered3
I
II
1—
• l NC
Grass sod
Cumulative
infiltration
in 60 min.
4.5
5.4
Mean
1.9
4.9
I
II
8.1
6.1
8.5
10.1
Mean
7.1
9.3
I
II
1.0
0.5
6.4
5.1
Mean
0.8
5.8
As described in the text, these values do not represent
true infiltration rates owing to excessive lateral movement
of water around the plot frame.
are significant at the 0.05 level.
Cumulative infiltration
has been used widely (see literature review) to relate infil­
tration to antecedent moisture by regression techniques.
A comparison of cumulative infiltration means for the
"moist" grass surface, the screen covered, and the bare sur­
face conditions shows that tillage increased the total water
intake in a 60 minute period 2.0 cm. above that for the grass
surface, but when the tilled surface was exposed to rainfall
100
impact no increased intake was realized.
Also, the sealing
effect is seen to reduce the equilibrium infiltration rate to
about half that on the untilled or protected surfaces.
ferences are significant at the 0.25 level.
Dif­
Similar conclu­
sions apply to the Grundy soil for which data are given in
Table 5.
The tillage effect on infiltration is the result of
an increase in total porosity with a concomitant decrease in
volumetric moisture content.
Tillage increased the pore space
from 51 to 72 percent on the Ida soil, and from 52 to 65 per­
cent on the Grundy.
Calculated Infiltration
Comparison of calculated and measured infiltration
Infiltration rates calculated from both desorption and
absorption data are compared with field measured rates in
Figure 13.
The calculated rates fit the measured rates best
for the "wet" condition and poorest for the "dry" condition.
Desorption data resulted in the best estimates for all times
on the "dry" soil and for times less than 20 minutes on the
"moist" soil, while for times greater than 20 minutes in the
latter case, the absorption data gave the best prediction.
The reason for the low infiltration rates obtained with
absorption data is not apparent, but is probably the result
of using an approximate curve to obtain 0 values for the
101
0-15 cm. depth (Figure 6).
As mentioned previously, the
absorption points in Figure 6 appear to lie to the left of the
desorption points, so that D values for a given moisture con­
tent may be 2 to 3 times greater for absorption than for
desorption.
In the computer program K is given by Equation
(20), i.e. K = D(A0/z# ), so that if the value of D used for
absorption were too small, then K would also be too small.
K values in Figure 7 for the 0-15 cm. depth and calculated
from absorption data indicate that an error was introduced
when the same 0(B) relation was used for both desorption and
absorption calculations.
Since the infiltration rate is
greatly influenced by the capillary conductivity (Equation
(23)), the use of low K values could be expected to reduce
calculated intake rates.
The erratic infiltration rates for the dry condition
(desorption data) in Figure 13 are probably the result of the
extreme moisture potential differences and conductivity dif­
ferences between the wet and dry soil at the wetting front.
The value of
j in (23) is dependent on the average mois­
ture content in the first depth increment, which for small
times is quite low on the initially dry soil.
Hence
.
will also have a low value for small times, thereby causing
the initially low infiltration rate on the dry soil.
The use
of smaller depth increments would have improved the calculated
results.
Figure 13.
Comparison of measured and calculated infil­
tration rates for Ida silt loam at three
antecedent moisture levels, using both de­
sorption and absorption data
103
DRY
\
• xe
o^oO ~s -o..o..
_c~
o—0--0
FIELD
0
10
o — o-
MOIST
THEORY
r
o
• DES.
° AGS.
r—n——r—~t—t~—
20
30
40
50
TIME (MINUTES)
60
70
Figure 14.
Comparison of measured and calculated
cumulative infiltration for Ida silt loam
at three antecedent moisture levels, using
both desorption and absorption data
105
5
DRY
4
3
- o*
--o-
2
^O-O-
:i
1
?
o
0
FIELD
4
MOIST
THEORY
<
cr
i-
3
2
.A
• DES.
O ABS.
-•*"
1
eoLU
>
0
H
<
4
WET
3
O
2
I
/•
0
0
10
20
30
40
TIME (MINUTES)
50 60
70
106
Although the measured infiltration rates in Figure 13
appear higher at times less than 10 minutes than the calcu­
lated rates (desorption data), the actual infiltration rate
curve is probably intermediate to the measured and calculated
curves at small times.
This discrepancy between measured and
calculated rates can be explained in part by the method of
field measurement.
Since spray application rates were less
than the potential infiltration rate until runoff occurred,
the measured rate curves shown in Figure 13 are steeper at
times between 5 and 10 minutes than they would have been if
water application at all times had been in excess of the in­
filtration rate.
The relatively inaccurate prediction of infiltration
rates at small times on the dry soil is reflected in the
cumulative curves in Figure 14.
However, the agreement be­
tween measured and calculated curves for the moist and wet
soil is considered excellent.
These results indicate (1)
that the capillary conductivity values obtained by the out­
flow procedure are quite accurate, (2) that Hanks and Bowers'
(1962a) numerical method gives valid estimates of infiltra­
tion, and (3) that moisture diffusion theory can be a useful
tool in the study of practical field problems.
107
Calculated infiltration for various
antecedent moisture profiles
The calculated curves shown in Figures 13 and 14 are
presented together in Figure 15.
In addition, data points
for a hypothetical "uniform wet" condition are shown for
comparison with the "wet" curves.
The moisture profile
curves in the upper-right-hand corner of the figure show the
relative amount of water storage capacity initially available
in the soil for each of the antecedent moisture conditions.
The "uniform wet" condition, although not too realistic
in view of the different x|>(9) relations for the two layers,
was included to predict whether displacement of initial water
in a uniformly wet profile would increase infiltration above
that on a profile with wet surface and drier subsurface con­
ditions.
Considering the expression due to Philip (1957b),
Equation (14), it is conceivable that the Kn term could ac­
count for a greater fraction of cumulative infiltration on
the "uniform wet" than on the "wet" soil.
However, Kn values
corresponding to moisture contents at or below field capacity
are not usually great enough to result in much displacement
of initial water.
If the conductivities corresponding to the
moisture contents at 30 cm. depth on the moisture profile
curves for the "wet" and "uniform wet" conditions (Figure 15)
are used for the values of
for each case, the K's (taken
from Figure 7) are respectively, K.(27%) = 1.5x10
4cm.
per hr.
Figure 15.
Calculated infiltration for Ida silt loam at four antecedent
moisture levels (desorption data)
MOISTURE %
(BY VOL.)
10
30
50
10 -
or
DRY
MOIST
WET
UNIFORM WET
S 10-
n
o
vO
0
10
30
40
TIME (MINUTES)
20
50
60
110
and K(35%) = 5.5x10 ^ cm. per hr.
Although the K at the 35
percent moisture content is 36 times that at 27 percent mois­
ture, both of these antecedent K values are small compared to
the K for saturation (Kq = 6x10
cm. per hr.), so that the
Kn term appears insignificant in both cases.
In addition,
the cumulative infiltration for both conditions is seen in
Figure 15 to be more than 3 cm. for a 60 minute period, so
_n
the 5.5x10
cm. flow given by the K t term in (14) is
negligible in this case.
Thus, the reason for the slightly
greater calculated cumulative infiltration on the "uniform
wet" than on the "wet" soil is obscure.
The equivalence of
calculated infiltration rates indicates little effect of the
higher moisture content at greater depths.
Calculated moisture profiles corresponding to the infil­
tration curves in Figure 15 are shown in Figure 16.
Three
points are noteworthy:
(a) The wetting front is most distinct for the ini­
tially dry soil and becomes more gradual with
increasing initial moisture contents.
(b)
The rate of advance of the wet front is propor­
tional to the antecedent moisture level.
(c)
The moisture content of the upper 10 cm. approaches
saturation (55 percent) for times greater than 60
minutes, especially on the wet soils.
Considerable attention has been given in previous studies to
moisture profiles and pressure gradients in the soil as re­
lated to infiltration rates.
The effects of initial moisture
Figure 16.
Calculated moisture profiles with time
during infiltration for Ida silt loam at
four antecedent moisture levels (desorption
data)
112
MOISTURE CONTENT, (% BY VOL.)
40
60 0
40
60
0
32 i/
10 -
20 -
30 -
MOIST
o
. 40
i
l-
20
0
O
t
40
o
H
QLD
60 0
20
40
A
\
9r
10 -
/1
22."^
y
38.^.56y
20 -
.7
/
30 40 -
WET
I
11
J
•1i
j
UN FORM i
60
113
content on the abruptness of the boundary between wet and dry
soil at the wetting front and on the rate of advance of the
wet front have been noted in laboratory studies and are pre­
dicted by theory.
Colman and Bodman (1944), in comparing the
rate of wet front advance on dry and moist soils, concluded
that the increased rate of downward penetration of the wet
front in the initially moist soil was the result of downward
displacement of initial moisture.
Philip (I957f) in a detailed
theoretical analysis found the velocity of wet front advance
to be given by the expression u = (Kq - Kn)/(0Q - ©n) as t in­
creases without limit.
A graph of u versus 6n, the initial
moisture content, shows that u increases rapidly as @n in­
creases.
In addition to their observations on wet front advance,
Bodman and Colman (1944) distinguished a number of zones in
the soil column during infiltration.
These zones have been
further defined by Philip (1957d) as follows:
(a)
the saturated zone near the surface, about 1.5 cm.
deep;
(b)
the transition zone, a region of rapid decrease of
moisture content extending to a depth of about 5
cm. from the surface;
(c)
the transmission zone, a region underlying (b) in
which the moisture content varies only slowly with
both x and t and which lengthens as infiltration
proceeds;
(d)
the wetting zone, in which moisture content changes
rapidly both with respect to x and t;
114
(e)
the "wet" front in which a very steep moisture
gradient represents the visible limit of moisture
penetration into the column.
The calculated profiles in Figure 16 do not predict zones (b)
and (c), but rather show a zone of near saturation from the
surface to the wetting zone.
Although Philip (1957b) obtained
calculated results similar to those in Figure 16, he concluded
that the "dog-legs" in the moisture profile curves of Bodman
and Colman (1944) represented a real phenomenon.
Previous
work on the Ida soil by Nielsen et al. (1961), in which mea­
sured soil moisture profiles were compared with profiles
calculated by Philip's numerical procedure, also showed a
steeper moisture gradient in the transmission zone of field
measured curves than on curves calculated from diffusion
theory.
The failure of Philip's analysis to account for the
transition zone appears to be the result of assuming a unique
aj)(6) relation (Philip, 1957d).
Moisture content and moisture
potential measurements of Miller and Richard (1952) suggest a
non-unique i|)(Q) relation for small distances from the surface,
and support the presence of a "transition" zone.
In a later
paper, Philip (1958b) defines a "tension-saturated" zone near
the surface in which the moisture content is that of satura­
tion but the value of i|) is negative, so that i|)(0) is not
unique in this zone.
The tension-saturated condition is,
according to Philip, the result of a non-zero air entry value
at the surface of the soil during infiltration.
Thus it
115
appears that the curves in Figure 16 may be somewhat in error
with respect to the transition and transmission zones, but
otherwise, the predicted curves are in accord with measured
profiles.
Moisture suctions corresponding to the "uniform wet" and
"dry" moisture content curves of Figure 16 are summarized in
Table 6.
The horizontal dashed lines are used to arbitrarily
separate what appear to be the "transmission" and "wetting"
zones.
As would be expected, the suction gradients at the
wetting front are much steeper on the dry soil than on the
wet.
It is probable that the magnitudes of the suction
gradients across the wetting front and in the wetting zone
largely determine the infiltration rate when the soil surface
is not subject to surface sealing.
Taylor and Heuser (1953)
concluded from measurements of moisture potential on field
cores that the suction gradient at the wetting front and in
the wetting zone were more important than transmission zone
gradients in determining infiltration rates.
If the suction
gradients at or near the wetting front do represent the
principal force causing water to move into the soil, then
one might expect measured "equilibrium rates" obtained after
a period of 2 or 3 hours to be greater for an initially dry
soil than for an initially wet soil.
then, appear reasonable.
The results in Table 4,
Although infiltration theory pre­
dicts that infiltration rate curves for a number of antecedent
116
Table 6.
,
Depth
cm.
Calculated soil water suction (cm. water) for diff
of infiltration with the "uniform wet" and "dry" c
to moisture contents in Figure 16a
Uniform wet
Elapsed time, min.
0.6
9.2
22.5
38.6
56.5
6.4
0- 2
207.9
1.5
0.7
0.4
0.2
125.1
2- 4
350.0
36.1
1.4
0.7
0.4
17714.4
4- 6
229.1
16.2
1.0
0.6
6- 8
347.1
54.5
3.6
0.9
241.6
22.8
1.2
10-12
61.6
4.0
12-14
144.1
25.6
14-16
177.5
56.6
8-10
16-18
96.8
18-20
146.4
20-22
173.2
For each time the suction at the greatest depth shown
soil at that depth at zero time. Suctions at greater depths
pertinent values.
^The horizontal dashed lines in the table designate th<
the "transmission" and "wetting" zones.
water) for different depths and times after the beginning
et" and "dry" conditions with suction values corresponding
1
1
Dry
Elapsed time, min.
37.2
25.3
6.4
14.8
51.7
67.5
125.1
9.2
0.8
+0.4
0.9
0.6
17714.4
408.9
15.3
2.0
1.8
1.3
17991.6
1715.0
25.5
17.0
2.4
16000.0
7504.2
64.9
25.5
12940.4
161.1
13653.0
;st depth shown is relatively close to the suction of the
b greater depths are excluded to facilitate reading of
Le designate the approximate depth of the boundary between
117
moisture contents will be asymptotic to the rate corresponding
to the saturated hydraulic conductivity (Figure lb), very
large times may be required for the transmission zone to be­
come so extended that it exhibits a unit hydraulic gradient.
When this occurs the wetting front is so far from the soil
surface (on initially dry soils) that the force resulting from
the large suction gradient at the wetting front is no longer
transmitted to the surface.
Such long periods seldom occur
with simulated or natural rainfall, hence a difference in
11 equilibrium"
rates for initially dry and wet soils can be
expected.
Suctions for the transmission zone in Table 6 are lower
than the 5-20 cm. water measured by Taylor and Heuser (1953),
Miller and Richard (1952), and Marshall and Stirk (1949).
The
low calculated suctions would be expected since the calculated
moisture profiles have correspondingly high values.
Calculated infiltration with an approximate uniform profile
The previous results indicate a surprisingly accurate
prediction of antecedent moisture effects from calculations
utilizing diffusion theory and laboratory measurements on
field cores.
The laboratory data used in the calculations
were obtained from distinct depths in the soil profile, and
therefore, could only approximate the unsaturated moisture
flow characteristics of the true profile,
It is of interest
118
to consider the necessity of accounting for the flow charac­
teristics of horizons below the plow layer in attempting to
predict infiltration and water movement in the profile.
Re­
sults obtained by Colman and Bodman (1944), Miller and Gardner
(1962) and Hanks and Bowers (1962a) indicate that if the sur­
face horizon is more permeable than the underlying horizon,
the latter may essentially control the infiltration process.
This is especially true on initially wet soils or for long
periods of infiltration.
If the surface horizon is suf­
ficiently conductive relative to the underlying soil, satura­
tion may occur at the boundary of the two horizons, and posi­
tive pressures may develop just as if the upper limit of the
lower horizon were at the surface and subjected to a positive
head.
However, if the surface soil is initially dry, lower
horizons would, in general, have little influence on infil­
tration until sufficient time had elapsed for the wetting
front to reach the less conductive layer.
A possible excep­
tion is that described by Miller (1959) in which a dense
underlying layer limits air movement so that positive air
pressures develop ahead of the wetting front, causing a reduc­
tion in the water intake rate.
On soils in which the under­
lying horizon is more conductive than the surface horizon,
the flow characteristics of the surface soil would be expected
to dominate, especially when the surface layer is sufficiently
deep (Hanks and Bowers, 1962b).
Although this situation is
119
not as common as the case of decreasing permeability with
depth, it is characteristic of certain Regosols, such as the
Ida in the present study, in which little eluviation of col­
loids has occurred.
A limiting surface conductivity also
occurs when the immediate soil surface has been rendered
impermeable by dispersion due to high sodium saturation, or
as a result of surface sealing by rainfall impact, as dis­
cussed previously.
The prediction of infiltration from laboratory measure­
ments would obviously be simplified if the moisture flow
characteristics of only the surface horizon had to be evalu­
ated.
To determine the feasibility of such a procedure, in­
filtration calculations were accomplished using conductivity
and moisture retention data from the 0-15 cm. depth of the
Ida soil as an approximation of the flow characteristics of
the entire profile.
The data points in Figure 17 corre­
sponding to the approximate single-horizon profile fall very
near to the infiltration curves obtained using the twohorizon approximation.
Thus, in this case, only the sur­
face 15 cm. needed to be considered.
In summary, it appears that valid prediction of infil­
tration may be accomplished using conductivity and moisture
retention data for the surface horizon only, providing the
lower horizons are more conductive to water.
Also, evidence
in the literature indicates that surface horizon data may
Figure 17.
Calculated infiltration for Ida silt loam with two simulated
profiles, the "approximate" (1 horizon) and the "real" (2
horizons) using desorption data
cr
PROFILE
10 -
u 8 -
APPROX
o
•
-5
I.R.
C.I.
-4
REAL —
o
«- 3 uto
ZD
O
0
10
20
30
40
TIME (MINUTES)
50
60
122
adequately describe infiltration for short periods on soils
which have less conductive horizons at greater depths if the
antecedent moisture level is sufficiently low.
123
SUMMARY AND CONCLUSIONS
The influence of antecedent soil moisture on water in­
filtration was studied in the field and laboratory using the
following approach:
(a) The antecedent moisture effect on water infiltra­
tion was estimated by use of moisture diffusion
theory employing the numerical procedure of Hanks
and Bowers (1962a) and utilizing laboratory mea­
surements of capillary conductivity and moisture
retention.
(b) Field measurements of infiltration were made on Ida
silt loam at three antecedent moisture levels to
check the validity of theory calculations. Other
infiltration measurements on both the Ida and
Grundy soils under different surface conditions
provided an evaluation of the effects of tillage,
surface sealing, and soil morphological differences
relative to the effect of antecedent moisture.
(c)
A laboratory rainfall simulator was used to deter­
mine the influence of initial moisture content of
Ida and Grundy soil aggregates on surface sealing
by rainfall impact.
Calculated infiltration rates were in close agreement with
field measured rates, both indicating a marked effect of
antecedent moisture on infiltration.
The best prediction was
accomplished on the initially "wet" soil, with the calculated
rate for the "dry" condition being quite erratic and slightly
below the measured infiltration rate.
The irregularity of
the rate curve for the dry soil is probably the result of
extreme moisture potential and conductivity differences be­
tween the wet and dry soil at the wetting front.
Moisture
124
retention data for desorption resulted in better infiltration
prediction than did the absorption data.
The relatively inac­
curate calculated rates in the latter case may have resulted
from using inappropriate diffusivity values in the calcula­
tions.
Calculated infiltration for two antecedent moisture con­
ditions, which were equally wet in the surface 4 cm. and
which differed in moisture content by 8 percent at the 30 cm.
depth, indicated no significant difference in water intake.
Displacement of initial profile water by infiltrating water
on the soil which was wettest (35 percent water) at 30 cm.
depth was shown to contribute little to the cumulative infil­
tration.
Calculated moisture profile curves are similar in
shape to those obtained by Philip (1957b), who used a dif­
ferent method to solve the partial differential equation for
moisture diffusion.
The curves demonstrate a "wetting zone"
and "wet front", but do not predict the "saturated", "transi­
tion", and "transmission" zones observed by Bodman and Colman
(1944) since the depths which might be expected to evidence
the latter three zones show predicted moisture contents near
saturation.
Accurate prediction of moisture profiles may in­
volve the use of a non-unique t|>(0) relation near the soil
surface to account for the "tension-saturated" zone suggested
by Philip (1958b).
The calculated moisture profiles do pre­
dict an abrupt wetting front on the initially dry soil and a
125
diffuse front on the wet soil, the wetting front becoming
more diffuse with time in both cases.
In addition, the rate
of advance of the wet front is proportional to the antecedent
moisture level, in accord with published experimental data.
Calculated suction differentials at the wetting front
indicate that the high equilibrium infiltration rates mea­
sured on the dry soil in the field, relative to those mea­
sured on initially wet soil, were the result of the high
suction gradient on the dry soil.
Infiltration calculations, based on conductivity and
moisture retention data from the surface horizon only, gave
results which agreed favorably with those obtained when data
for a "two-layer profile" were used.
The surface-soil ap­
proximation appears to be adequate in cases where the under­
lying horizon is more conductive than the surface soil, and
in cases where the lower horizon is less conductive if the
antecedent moisture level is sufficiently low.
Capillary conductivity data for the Ida soil, obtained
by the outflow method, show definite differences in the
moisture flow characteristics of soil from the 0-15 cm., 1522.5 cm., and 22.5-30 cm., depths, with excellent duplication
of results being obtained on the latter two depths.
Con­
ductivity measurements for both desorption and absorption
Indicate a unique Kl©) relation for the 15-30 cm. depth even
though hysteresis in the ij)(©) relation was evident.
On the
126
other hand, K values for the 0-15 cm. depth measured during
absorption are greater at a given moisture content than
desorption K values, indicating a non-unique K(9) relation
for this depth increment where hysteresis in i|)(9) is more
pronounced.
Field infiltration rates measured on the Ida and Grundy
soils show that in some cases antecedent moisture differences
on a given soil may influence infiltration rates as much as
tillage, surface sealing, or profile differences between
soils.
Tillage not only changes the volume and geometry of
soil pores, but at the same time decreases the volumetric
moisture content in the surface soil.
Surface sealing
diminishes the effect of antecedent moisture on infiltration
since the hydraulic conductivity of the immediate soil sur­
face limits water flow into the soil and does not allow suc­
tion gradients to control the rate of infiltration.
Surface
sealing experiments with a laboratory rainfall simulator
showed that the initial moisture content of soil aggregates
did affect the rate of surface sealing, but the moisture
effect on sealing was different on the two soils.
The Ida
aggregates were most stable when initially "dry", while the
"wet" Grundy aggregates resisted sealing more than those
which were "moist" or "dry".
A difference in the nature of
bonds between particles is probably responsible for the
measured difference on the two soils.
127
The excellent agreement of theoretical and measured in­
filtration in the present study should encourage the applica­
tion of moisture diffusion theory to a variety of infiltration
problems.
Various infiltration characteristics calculated
for a wide range of conditions can be checked against field
results at a few points to establish the validity of theory
calculations.
In this way the theory can be used to broaden
our understanding of phenomena which have important practical
implications.
The K(9) and 4(9) relations are measurable variables of
physical significance in theoretical equations describing
water flow in phenomena such as water infiltration, evapora­
tion, and moisture movement to plant roots.
These useful
moisture flow properties have been neglected until recent
years owing to the difficulty of laboratory measurement of
capillary conductivity and the restrictive conditions required
to solve moisture flow equations.
The outflow procedure of
measuring capillary conductivity, as modified by Kunze and
Kirkham (1962), and the use of high speed computers to facili­
tate solution of flow equations, as exemplified by the method
of Hanks and Bowers (1962a), make possible a broader applica­
tion of the moisture diffusion concept to complex problems
which arise in the field.
128
LITERATURE CITED
Arend, J. L. and Horton, R. E. 1943. Some effects of rain
intensity, erosion, and sedimentation on infiltration
capacity. Soil Sci. Soc. Amer. Proc. 7: 82-89.
Ayers, H. P. and Wikramanayoke, V. E. A. 1958. The effect
of the water-storage capacity of the soil on mass
infiltration. Canada Jour. Soil Sci. 38: 44-48.
Baver, L. D. 1938. Ewald Wollny, a pioneer in soil and
water conservation research. Soil Sci. Soc. Amer. Proc.
3: 330-333.
Bertoni, Jose, Larson, W. E. and Shrader, W. D. 1958.
Determination of infiltration rates on Marshall silt
loam from runoff and rainfall records. Soil Sci. Soc.
Amer. Proc. 22: 571-574.
Bertrand, A. R. and Parr, J. F. 1961. Design and operation
of the Purdue sprinkling infiltrometer. Indiana Agr.
Expt. Sta. Res. Bui. 723.
Black, C. A. 1957. Laboratory methods of soil investigationssoil fertility. 3rd ed. (Mimeo.) Department of
Agronomy, Iowa State University of Science and Technology,
Ames, Iowa.
Bodman, G. B. and Colman, E. A. 1944. Moisture and energy
conditions during downward entry of water into soils.
Soil Sci. Soc. Amer. Proc. 8: 116-122.
Brakensiek, D. L. and Frevert, R. K. 1961. Analysis and
application of infiltrometer tests. Trans. Am. Soc.
Agric. Eng. 4: 75-77.
Browning, G. M. 1939. Volume changes of soils in relation
to their infiltration rates. Soil Sci. Soc. Amer. Proc.
4: 23-27.
Bruce, R. R. and Klute, A.
moisture diffusivity.
458-462.
1956. The measurement of soil
Soil Sci. Soc. Amer. Proc. 20:
Buchele, Wesley F. 1961. A power sampler of undisturbed
soils. Trans. Am. Soc. Agr. Eng. 4: 185-187, 191.
129
Buckingham, Edgar. 1907. Studies on the movement of soil
moisture. U. S. Bur. Soils Bui. 38.
Cernuda, C. F., Smith, R. M. and Vicente-Chandler, j. 1954.
Influence of initial soil moisture on resistance of
macro-aggregates to slaking and to water-drop impact.
Soil Sci. 77: 19-27.
Childs, E. C. 1936a. The transport of water through heavy
clay soils. I. Jour. Agr..Sci. 26: 114-127.
. 1936b.
soils. III.
The transport of water through heavy clay
Jour. Agr. Sci. 26: 527-545.
Colman, E. A. and Bodman, G. B. 1944. Moisture and energy
conditions during downward entry of water into moist and
layered soils. Soil Sci. Soc. Amer. Proc. 9: 3-11.
Diebold, C. H. 1951. Soil layers causing runoff from hardland wheat fields in Colorado and New Mexico. Jour.
Soil Water Conser. 6: 202-209.
Duley, F. L. 1939. Surface factors affecting the rate of
intake of water by soils. Soil Sci. Soc. Amer. Proc.
4: 60-64.
Ellison, W. D. 1945. Some effects of raindrops and surface
flow on soil erosion and infiltration. Trans. Am.
Geophys. Union 26: 415-429.
Evans, D. D. 1954. Effect of prewetting and incubation of
soil on aggregate analysis. Soil Sci. Soc. Amer. Proc.
18: 10-12.
Free, G. R., Browning, G. M. and Musgrave, G. W. 1940.
Relative infiltration and related physical character­
istics of certain soils. U. 5. Dept. Agric. Tech. Bui.
729.
Gardner, W. and Widtsoe, J. A. 1921. The movement of soil
moisture. Soil Sci. 11: 215-233.
Gardner, W. R. 1956. Calculation of capillary conductivity
from pressure plate.outflow data. Soil Sci. Soc. Amer.
Proc. 20: 317-320.
. 1959. Diffusivity of soil water during sorption
as affected by temperature. Soil Sci. Soc. Amer. Proc.
23: 406-407.
130
and Mayhugh, M. S. 1958. Solutions and tests of
the diffusion equation for the movement of water in
soil. Soil Sci. Soc. Amer. Proc. 22: 197-201.
Green, R. E., Larson, W. E., Amemiya, M. and Moldenhauer,
W. C. [1961]. Annual report of cooperative regional
projects, project NC-40, water infiltration into soils.
(Mimeo.) Department of Agronomy, Iowa State University,
Ames, Iowa.
Hanks, R. J. and Bowers, S. A. [1962a]. Numerical solution
of the moisture flow equation for infiltration in
layered soils. [To be published in Soil Sci. Soc. Amer.
Proc.].
and
. [1962b]. Estimation of infiltration
rates from soil moisture properties. [An annual report
to the regional committee on infiltration]. (Mimeo.)
Department of Agronomy, Kansas State University,
Manhattan, Kansas.
Hansen, V, E. 1955. Infiltration and soil wgter movement
during irrigation. Soil Sci. 79: 93-105.
Hodgman, C. D., ed. 1960. Handbook of chemistry and physics.
41st ed. Chemical Rubber Publishing Co., Cleveland,
Ohio.
Horton, R. E. 1933. The role of infiltration in the
hydrologie cycle. Trans. Am. Geophys. Union 14: 446-460.
. 1940. An approach toward a physical interpretation
of infiltration-capacity. Soil Sci. Soc. Amer. Proc.
5: 399-417.
Jamison, V. C. and Thornton, J. Jr. 1961. Water intake rates
of a claypan soil from hydrograph analyses. Jour.
Geophys. Res. 66: 1855-1860.
Kilmer, V. J. and Alexander, L. T.
mechanical analyses of soils.
1949. Methods of making
Soil Sci. 68: 15-24.
Kirkham, Don and Feng, C. L. 1949. Some tests of the dif­
fusion theory and laws of capillary flow in soils.
Soil Sci. 67: 29-40.
Klute, A. 1952. A numerical method for solving the flow
equation for water in unsaturated materials. Soil Sci.
73: 105-116.
131
Kunze, R. J. and Kirkham, Don. [1962]. Simplified accounting
for membrane impedance in capillary conductivity deter­
minations. [To be published in Soil Sci. Soc. Amer.
Proc., Vol. 26].
Ligon, James T. and Johnson, Howard P. 1960. Infiltration
capacities of Fayette silt loam from analysis of
hydrologie data. Trans. Am. Soc. Agr. Eng. 3: 36-37.
Marshall, T. J. and Stirk, G. B. 1949. Pressure potential
of water moving downward into soil. Soil Sci. 68: 359370.
Mclntyre, 0. S. 1958. Permeability measurements of soil
crusts formed by raindrop impact. Soil Sci. 85: 185-189.
Miller, D. E. 1959. Effect of profile stratification and
other factors on water infiltration. Unpublished Ph. 0.
thesis. Library, Washington State University, Pullman,
Washington.
and Gardner, W. H. 1962. Water infiltration into
stratified soil. Soil Sci. Soc. Amer. Proc. 26: 115-119.
Miller, E. E. and Elrick, 0. E. 1958. Dynamic determination
of capillary conductivity extended for non-negligible
membrane impedance. Soil Sci. Soc. Amer. Proc. 22: 483486.
Miller, R. D. and Richard, F. 1952. Hydraulic gradients
during infiltration in soils. Soil Sci. Soc. Amer.
Proc. 16: 33-38.
Mutchler, C. K. and Moldenhauer, W. C. [1962]. An applicator
for a laboratory rainfall simulator. [To be published
in Agr. Eng.].
Neal, Jesse H. 1938. The effect of the degree of slope and
rainfall characteristics on runoff and soil erosion.
Missouri Agr. Expt. Sta. Res. Bui. 280.
Nielsen, D. R. and Biggar, J. W. 1961. Measuring capillary
conductivity. Soil Sci. 92: 192-193.
, Kirkham, Don and Perrier, E. R. 1960. Soil
capillary conductivity: comparison of measured and
calculated values. Soil Sci. Soc. Amer. Proc. 24: 157-
160.
132
,
and van Wijk, W. R. 1959. Measuring
water stored temporarily above the field capacity. Soil
Sci. Soc. Amer. Proc. 23: 408-409.
,
and
. 1961. Diffusion equation
calculations of field soil water infiltration profiles.
Soil Sci. Soc. Amer. Proc. 25: 165-168.
Parr, J. P. and Bertrand, A. R. 1960. Water infiltration
into soils. Adv. in Agron. 12: 311-363.
Philip, J. R. 1955. The concept of diffusion applied to
soil water. Natl. Acad. Science Proc. (India) 24: 93104.
. 1957a. Numerical solution of equations of the
diffusion type with diffusivity concentrâtion-dependent
II. Aust. Jour. Phys. 10: 29-42.
. 1957b. The theory of infiltration I. The
infiltration equation and its solution. Soil Sci. 83:
345-357.
. 1957c. The theory of infiltration II.
profile at infinity. Soil Sci. 83: 435-448.
The
. 1957d. The theory of infiltration III. Moisture
profiles and relation to experiment. Soil Sci. 84: 163178.
1957e. The theory of infiltration IV.
Sorptivity and algebraic infiltration equations.
Sci. 84: 257-264.
Soil
. 1957f. The theory of infiltration V. The
influence of the initial moisture content. Soil Sci. >
84: 329-339.
. 1958a. The theory of infiltration VI. Effect
of water depth oyer soil. Soil Sci. 85: 278-286.
1958b. The theory of infiltration
Sci. 85: 333-337.
VII.
Soil
Reinhart, K. G. and Taylor, R. E. 1954. Infiltration and
available water storage capacity in the soil. Trans.
Am. Geophys. Union 35: 791-795.
133
Richards, L. A. 1931. Capillary conduction of liquid through
porous media. Physics 1: 318-333.
. 1952. Report of the subcommittee on permeability
and infiltration, committee on terminology. Soil Sci.
Amer. Proc. 16: 85-88.
Robinson, D. 0. and Page, J. B. 1951. Soil aggregate
stability. Soil Sci. Soc. Amer. Proc. 15: 25-29.
Schiff, L. and Dreibelbis, F. R. 1949. Preliminary studies
on soil permeability and its application. Trans. Am.
Geophys. Union 30: 759-766.
Sokolnikoff, I. S. and Redheffer, R. M.
of physics and modern engineering.
Co., Inc., New York, N. Y.
1958. Mathematics
McGraw-Hill Book
Taylor, 5. A. and Heuser, N. C. 1953. Water entry and down­
ward movement in undisturbed soil cores. Soil Sci. Soc.
Amer. Proc. 17: 195-201.
Thames, J. L. and Ursic, S. J. I960. Runoff as a function
of moisture and storage capacity. Jour Geophys. Res. 65:
651-654.
Tisdall, A. L. 1951. Antecedent soil moisture and its rela­
tion to infiltration. Aust. Jour. Agr. Res. 2: 342-348.
Watson, K. K. 1959. A note on the field use of a theoreti­
cally derived infiltration equation. Jour. Geophys.
Res. 64: 1611-1615.
Yoder, R. E. 1936. A direct method of aggregate analysis of
soils and a study of the physical nature of erosion
losses. Jour. Am. Soc. Agron. 28: 337-351.
Youngs, E. G. 1957. Moisture profiles during vertical
infiltration. Soil Sci. 84: 283-290.
134
ACKNOWLEDGMENTS
The author wishes to express special appreciation to Dr.
W. E. Larson for his counsel during the study and assistance
in preparation of the manuscript.
Appreciation is also extended to Dr. M. Amemiya, Dr.
W. C. Moldenhauer, and Dr. R. J. Hanks for their assistance
in various phases of the study.
The description of soil
profiles by E. C. A. Runge and 0. Lockridge is also appre­
ciated.
The author is truly thankful for the assistance of his
wife, Mary, in typing the preliminary draft and for much
encouragement throughout the graduate program.
Above all, acknowledgment is given to the One who has
made Himself known through His Son and has caused the author
to be especially aware of His guidance during the period of
graduate study.
135
APPENDIX A
Table 7.
Profile description of Ida silt loam at the loca­
tion3 of field infiltration measurements
Depth, cm.
0 - 1 6
Horizon
Morphology
Ap
Dark brown (10YR 3/3, moist) very
friable silt loam; weak fine granular
structure; common lime concretions
(less than 1/2 cm.)
16 -
34
C,,
Dark yellowish brown (10YR 4/4, moist)
friable silt loam; weak medium subangular blocky to massive; common lime
concretions (less than 1/2 cm.)
34 -
51
C10
Dark yellowish brown to olive brown
(1Y 4/4, moist) friable silt loam; weak
medium subangular blocky to massive;
common lime concretions (less than 1/2
cm.)
51 -
66
C01
Dark yellowish brown to olive brown
(1Y 4.4/4, moist) friable silt loam;
massive; few lime concretions
66 - 122
CUr,
Yellowish brown to light olive brown
(1Y 5/4, moist) few fine faint strong
brown (7.5YR 5/8, dry) mottles;
friable silt loam; massive; very few
lime concretions
^Location:
890 feet south and 160 feet east from the NW
corner of NW%, NE%, Section 27, T84N, R43W,
Monona County, Iowa (Western Iowa Experi­
mental Farm).
136
Table 8. Profile description of Grundy silty clay loam at
the locationaof field infiltration measurements
Depth, cm.
Horizon
Morphology
0 -
20
A-,
Black (10YR 2/1, moist) friable silt
loam to light silty clay loam; dry
matrix color (10 YR 4/1); weak to
moderate fine granular structure
(slightly compacted)
20 -
33
A-
Mixed very dark gray (10Y 3/1, moist)
and dark gray (10YR 4/1, moist)
friable light to medium silt clay loam
with few fine dark grayish brown (10YR
4/2, moist) mottles; dry matrix color
gray (10YR 5/1); moderate very fine to
fine granular structure; few fine pores
33 -
48
Bp,
Dark gray (10YR 4/1, moist) firm light
silt clay with many fine yellowish
brown (10YR 5/6 and 5/8, moist) mot­
tles, and few fine very dark gray
(10YR 3/1, moist) mottles ; dry color
grayish brown (10YR 5/2); crushed
color (10YR 4.5/3, moist); moderate
very fine subangular blocky structure;
thick discontinuous clay films on ped
surfaces; few fine pores ; few 1-2 mm.
Mn-Fe concretions
48 -
64
Bpp
Dark gray brown (10YR 4/2, moist) firm
medium silty clay with common medium
yellowish brown (10YR 5/6 and 5/8,
moist) mottles and few medium black
and very dark gray (10YR 2/1 and 3/1,
moist) mottles; dry color light
brownish gray (10YR 6/2); crushed
color same as matrix; moderate to
^Location:
15 feet west and 20 feet north of NE corner
of runoff plots on the Grundy-Shelby Experi­
mental Farm. NW%}, SW%; T70N, R2SW, Section
35, Ringgold County, Iowa.
137
Table 8. (Continued)
Depth, cm.
Horizon
Morphology
strong fine subangular blocky struc­
ture; thick discontinuous clayfilms on
ped surfaces; common 1-3 mm. Mn-Fe
concretions; common fine pores
64 -
86
86 - 107
107 - 150
B31
B32C1
cb
Grayish brown (2.5Y 5/2, moist) firm
light silty clay with common medium
yellowish brown (lOYR 5/6 and 5/8,
moist) mottles and few medium dark
brown to brown (7.5YR 4/4, moist) mot­
tles; few very dark gray (10YR 3/1,
moist) vertical streaks ; crushed color
dark grayish brown to brown (10YR
4.5/2.5, moist); moderate fine to
medium subangular blocky to angular
blocky structure with long vertical
cleavage planes ; thin clay films on
moist peds; common 1-3 mm. concretions ;
common fine pores
Grayish brown (2.5Y 5/2, moist) firm
silty clay loam with common medium
dark brown to brown (7.5YR 4/4, moist)
mottles and common fine yellowish
brown (10YR 5/6 and 5/8, moist) mottles;
crushed color brown to yellowish brown
(10YR 5/3.5); weak medium subangular
blocky structure with long vertical
cleavage planes ; common 1-3 mm. concre­
tions ; many fine pores
Grayish brown (2.5Y 5/2, moist) firm
silty clay loam with common medium
dark brown to brown (7.5YR 4/4, moist)
mottles ; massive; common 1-3 mm. con­
cretions; many fine pores
^This horizon description is a composite taken from
other Grundy descriptions.
138
APPENDIX B
Table 9.
Measured values of 9, D, and K at various suctions, x|) , for both desorption and absorption on Ida silt loam cores from the 0-7.5 cm. depth
(measured on ceramic plate)
t|)
9
Unit No. 1
D
cm. water
%
cm.^/hr.
K
10
4
cm./hr.
x|>
9
Unit No. 5
D
cm. water
%
cm.^/hr.
(Desorption)
K
10
4
cm./hr.
(Desorption)
25
52.8
0.60
4.03
25
51.8
5.72
46.18
60
50.4
1.75
13.18
60
49.0
10.87
83.93
103
47.2
1.91
11.74
103
45.7
6.29
38.91
155
44.0
1.38
7.54
155
5.10
19.70
207
41.2
1.53
6.58
207
42.5
40.5
3.49
8.78
310
36.7
0.45
1.14
310
37.9
2.60
4.89
506
31.8
0.28
0.35
811
28.1
0.14
0.08
506
34.2
2.49
1.48
1240
25.5
0.03
0.01
310
35.4
2.30
2.68
207
36.5
1.31
2.15
155
37.4
1.32
3.58
100
38.8
7.38
33.91
60
40.7
0.39
3.48
25
43.8
2067
,
;
»
22.3
i
-
(Absorption)
-
-
Table 10.
Measured values of 9, D, and K at various suctions, t|), for both
desorption and absorption on Ida silt loam cores from the 0-7.5 cm.
depth (measured on millipore)
Unit No. 1
4
cm. water
9
%
Unit No. 2
D
O
cm. /hr.
K
10
4
6
cm./hr.
%
D
n
cm. /hr.
K
10' 4 cm./hr.
(Desorption)
0
53.1
105.0
1836
52.9
145.8
1320
10
51.3
336.6
1998
52.0
291.6
1254
27
50.3
174.6
600
51.2
109.2
702
45
49.8
64.8
302
50.3
91.8
714
70
48.6
4.4
47.3
48.3
19.0
224
100
45.4
8.2
41.3
44.8
28.5
200
.
(Absorption)
130
43.9
-
-
42.7
14.2
69
100
44.5
-
-
44.2
25.7
85
70
45.3
-
-
45.2
-
-
45
46.4
-
-
46.7
-
-
27
47.0
-
-
47.9
-
-
12
47.9
-
-
49.6
-
-
aD and K data are not included for absorption measurements which did not fit
the theory curves.
Table 11.
Measured values of 9, D, and K at various suctions, t|> , for both desorp­
tion and absorption on Ida silt loam cores from the 7.5-15 cm. depth
(measured on ceramic plate)
o|>
0
Unit No. 2
D
cm. water
%
cm.^/hr.
K
10
4
cm./hr.
i|)
9
Unit No. 6
D
cm. water
%
cm.^/hr.
(Desorption)
K
10
4
cm./hr,
(Desorption)
25
53.6
1.26
6.38
25
51.9
5.07
37.21
60
51.8
1.70
12.41
60
49.4
6.69
47.26
103
48.6
3.12
24.52
103
46.3
16.31
98.74
155
44.6
0.80
5.28
155
43.2
11.82
44.20
207
40.9
3.47
16.68
207
41.2
6.68
26.89
310
35.9
1.84
4.75
310
37.1
2.91
7.38
506
30.9
0.22
0.31
811
26.7
0.06
0.037
506
32.1
1.83
1.40
1240
23.9
0.02
0.009
310
33.6
2.00
3.94
2067
20.7
-
207
35.9
2.90
8.96
155
37.6
4.14
16.22
100
39.7
2.85
14.86
60
41.8
2.40
22.95
25
45.2
-
(Absorption)
-
-
Table 12.
4
cm. water
Measured values of 9, D, and K at various suctions, x|) , for both desorp­
tion and absorption on Ida silt loam cores from the 15-22.5 cm. depth
(measured on ceramic plate)
0
Unit No. 3
D
%
cm.^/hr.
K
10
4
cm./hr.
4»
cm. water
(Desorption)
9
%
Unit No. 7
0
p
cm. /hr.
K
10
4
cm./hr.
(Desorption)
25
51.3
0.99
14.5
25
51.8
7.84
119.6
60
46.2
10.03
127.7
60
46.5
22.14
241.1
103
40.7
14.19
131.7
103
41.8
31.52
262.1
155
35.9
8.69
56.0
155
37.5
123.3
207
32.6
12.15
47.0
207
34.4
20.82
11.96
310
28.6
2.67
5.6
310
30.4
4.33
9.4
506
24.5
0.26
0.29
811
21.1
19.0
0.087
0.043
506
26.2
1.53
1.6
0.021
0.007
310
28.2
2.11
5.4
16.4
-
-
207
30.9
4.77
20.9
155
33.2
36.6
8.59
53.7
11.22
92.2
7.79
89.8
1240
2067
46.3
(Absorption)
100
60
25
39.9
44.2
-
-
Table 13.
'
i|)
cm. water
Measured values of 0, D, and K at various suctions, ^, for both
desorption and absorption on Ida silt loam cores from the 7.5-15 and
15-22.5 cm. depths (measured on millipore)
Unit No. 3
0
D
%
cm.^/hr.
10
4
K
0
cm./hr.
%
Unit No. 4
D
cm.^/hr.
10
K
4
cm./hr.
(Desorption)
0
56.4
-
10
54.6
455.5
27
45
53.3
51.4
120.0
36.5
70
46.7
17.5
100
42.9
18.6
52.6
51.5
-
3438
1476
690
225
145
-
1302
-
7806
3534
50.5
49.0
350
-
-
47.3
140
1192
44.7
65
584
(Absorption)
130
40.5
17.5
52.5
42.0
11.0
65
100
41.4
7.9
34.9
43.8
13.9
85
70
42.8
7.8
52.2
45.7
15.5
118,
45
44.8
4.4
46.2
47.6
-
-
27
46.0
-
-
49.0
-
-
12
48.8
-
-
51.1
-
-
Table 14.
Measured values of 0, D, and K at various suctions, 4, for both desorp­
tion and absorption on Ida silt loam cores from the 22.5-30 cm. depth
(measured on ceramic plate)
4
0
Unit No. 4
D
cm. water
%
cm.^/hr.
K
10
4
cm./hr.
x|)
0
Unit No. 8
D
cm. water
%
cm.^/hr.
(Desorption)
K
10
4
cm./hr.
(Desorption)
25
45.9
0.12
2.33
25
45.6
1.47
26.5
60
39.1
1.38
19.86
60
39.3
29.79
300.4
103
32.9
1.64
14.29
103
34.9
17.90
117.6
155
2.21
12.78
155
31.5
173.76
825.7
207
28/4
25.4
6.77
21.7
207
310
29.1
310
1.92
1.89
30.51
7.97
100.8
14.6
506
18.0
0.41
0.34
811
15.5
0.23
0.06
1240
14.4
-
2067
13.7
-
3.60
25.7
(Absorption)
506
22.1
2.43
2.4
-
310
24.0
8.9
-
207
26.6
3.52
6.61
25.4
155
28.6
13.94
79.6
100
31.8
20.02
141.6
60
34.6
7.98
84.7
25
38.6
-
-
Table 15. Corresponding values of 0, 4aks > ^des ' anc^ ^ used in theory calcula­
tions of infiltration on Ida silt loam with two approximate horizons
0
Horizon 1 (0-15 cm.)
^abs.
cm. water
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
18000
16000
14000
12000
10000
6500
4000
3000
2400
1900
1550
1300
1100
950
820
720
630
550
490
440
390
350
310
275
240
^des.
cm. water
20000
18000
16000
14000
12000
10000
8000
6000
4500
3500
2900
2300
1850
1550
1350
1150
980
850
750
660
590
530
480
430
390
p ^
cm. /hr.
0.00024
0.00033
0.00046
0.00064
0.00090
0.00125
0.00175
0.00245
0.0034
0.0046
0.0065
0.0089
0.012
0.017
0.023
0,032
0.045
0.064
0.088
0.120
0.170
0.22
0.32
0.45
0.62
Horizon 2 (>15 cm.)
^abs.
cm. water
12000
9000
7500
4100
2700
2000
1700
1350
1125
950
830
710
620
540
475
420
370
325
285
250
225
195
175
155
140
^des.
cm. water
16000
14000
12000
10000
6000
4000
2900
2200
1700
1350
1150
950
820
700
610
530
470
410
370
330
295
265
240
220
200
^ ^
cm. /hr.
0.00034
0.00054
0.00087
0.0014
0.0022
0.0035
0.0056
0.0090
0.014
0.022
0.036
0.057
0.092
0.145
0.235
0.370
0.600
0.950
1.50
2.45
3.85
5.20
7.00
9.00
11.5
Table 15. (Continued)
0
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
Horizon 1 (0-15 cm.)
^abs.
^des.
2°
cm. water
cm. water
cm. /hr.
215
190
160
135
115
95
73
57
44
34
27
19
13.5
9.4
6.3
4.0
2.6
1.7
0
350
320
290
270
250
230
205
185
170
155
140
125
110
93
80
65
50
30
20
3
0
0.86
1.20
1.65
2.30
3.20
4.50
6.20
8.50
11.0
16.5
23
31
40
52
68
90
118
158
215
320
600
Horizon 2 (>15 cm.)
^abs.
^des.
2°
cm. water
cm. water
cm. /hr.
125
110
96
. 82
70
59
49
40
33
26
19
14
9.5
5.8
3.3
1.8
0
180
165
150
135
125
115
105
97
88
79
71
63
55
48
40
32
19
5
0
14.0
17.0
21.0
25.0
30.0
34.5
40
46
54
62
74
90
115
145
185
250
520
1800
5800